HUMAN ECOLOGY, AGRICULTURAL INTENSIFICATION AND LANDSCAPE
TRANSFORMATION AT THE ANCIENT MAYA POLITY OF
UXBENKÁ, SOUTHERN BELIZE
by
BRENDAN JAMES CULLETON
A DISSERTATION
Presented to the Department of Anthropology
and the Graduate School of the University of Oregon in partial fulfillment of the requirements
for the degree of Doctor of Philosophy
March 2012
ii
DISSERTATION APPROVAL PAGE Student: Brendan James Culleton Title: Human Ecology, Agricultural Intensification and Landscape Transformation at the Ancient Maya Polity of Uxbenká, Southern Belize This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Anthropology by: Douglas J. Kennett, Ph.D. Chairperson Jon M. Erlandson, Ph.D. Member Madonna L. Moss, Ph.D. Member Patrick Bartlein, Ph.D. Member Keith M. Prufer, Ph.D. Outside Member and Kimberly Andrews Espy Vice President for Research & Innovation/Dean of the
Graduate School Original approval signatures are on file with the University of Oregon Graduate School. Degree awarded March 2012
iii
© 2012 Brendan James Culleton
iv
DISSERTATION ABSTRACT Brendan James Culleton Doctor of Philosophy Department of Anthropology March 2012 Title: Human Ecology, Agricultural Intensification and Landscape Transformation at the
Ancient Maya Polity of Uxbenká, Southern Belize
Identifying connections between land use, population change, and natural and
human-induced environmental change in ancient societies provides insights into the
challenges we face today. This dissertation presents data from archaeological research at
the ancient Maya center of Uxbenká, Belize, integrating chronological,
geomorphological, and settlement data within an ecological framework to develop
methodological and theoretical tools to explore connections between social and
environmental change or stability during the Preclassic and Classic Period (~1000 BC to
AD 900).
High-precision AMS 14C dates from Uxbenká were integrated with stratigraphic
information within a Bayesian framework to generate a high-resolution chronology of
sociopolitical development and expansion in southern Belize. This chronology revises the
previous understanding of settlement and development of Classic Maya society at Uxbenká
and indicates specific areas of investigation to elucidate the Late and Terminal Classic
periods (AD 600-900) when the polity appears to disintegrate. A geoarchaeological record
of land use was developed and interpreted with respect to regional climatic and cultural
histories to track landscape transformations associated with human-environment
v
interactions at Uxbenká. The first documented episode of landscape instability (i.e.,
erosion) was associated with farmers colonizing the area. Later, landscape stability in the
site core parallels Classic Period urbanization (AD 300-900) when swidden agriculture was
likely restricted in the core. Another erosional event followed political disintegration as
farmers resumed cultivation in and around the abandoned city.
Maize yields derived from contemporary Maya farms in the area were used to
estimate the maximum population size of Uxbenká during its Classic Period peak. The
maximum sustainable population is estimated between 7500 and 13,000, including a
potential population of ~525 elites in the core, assuming low levels of agricultural
intensification. This accords well with the lack of archaeological evidence for intensive
land management during the Classic Period (e.g., terraces). An ecological model developed
using maize productivity and other environmental/social datasets largely predicts the
settlement pattern surrounding Uxbenká. Settlements in marginal areas may be evidence of
elite intra-polity competition during the Late Preclassic Period (ca. AD 1-300), though it is
possible that marginal areas were settled early as garrisons to mediate travel into the site
core.
This dissertation includes previously published and unpublished co-authored material.
vi
CURRICULUM VITAE NAME OF AUTHOR: Brendan James Culleton GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene University of California, Santa Cruz Cabrillo College, Aptos, CA DEGREES AWARDED: Doctor of Philosophy, Anthropology (Archaeology), 2012, University of Oregon Master of Science, Anthropology (Archaeology), 2006, University of Oregon Bachelor of Arts, Anthropology (Biological), 1996, University of California,
Santa Cruz AREAS OF SPECIAL INTEREST:
Human Responses to Quaternary Climate Change High-resolution Radiocarbon Correction and Calibration Modeling Diet through Stable Isotope Analyses Characterization of Marine and Freshwater Radiocarbon Reservoirs Use of Stable Isotopes to Model Hydrologic Systems and Determine Seasonality
PROFESSIONAL EXPERIENCE:
Research Technologist. Department of Anthropology, The Pennsylvania State University. 2011-present
Graduate Research Fellow. Department of Anthropology, University of Oregon.
2009-2011
Graduate Teaching Fellow. Department of Anthropology, University of Oregon. 2005-2006
Graduate Research Fellow. Department of Anthropology, University of Oregon.
2005
Archaeologist. Pacific Legacy, Incorporated. Santa Cruz, CA. 1999-2004
vii
GRANTS, AWARDS, AND HONORS:
Homer G. Barnett Fellowship for Course Design: Archaeological Sciences: Interdisciplinary Approaches to Understanding the Past, Department of Anthropology, University of Oregon, 2011. (Declined)
Educational Technology Grant ($5900): Tools for Integrated Landscape
Visualization and Ecological Analyses, Archaeometry Facility (with D.J. Kennett), College of Arts and Sciences and Department of Anthropology, University of Oregon, 2008.
National Science Foundation Doctoral Dissertation Improvement Grant: Human
Ecology, Agricultural Intensification and Landscape Transformation at the Ancient Maya Polity of Uxbenká, Southern Belize ($14,998; D.J. Kennett, P.I.), 2008.
Educational Technology Grant ($3150): Equipment for Bone and Shell Protein Isolation and Purification for Stable Isotope and Radiocarbon Analyses, Archaeometry Facility (with D.J. Kennett), College of Arts and Sciences and Department of Anthropology, University of Oregon, 2007.
National Science Foundation Graduate Research Fellowship, Cultural Responses
to Holocene Environmental Change in the Southern San Joaquin Valley, California, 2006-2009.
Educational Technology Grant ($6930): Acquisition of Equipment for 14C Sample
Preparation and Interpretation, Archaeometry Teaching Lab (with D.J. Kennett and J.M. Erlandson), College of Arts and Sciences and Department of Anthropology, University of Oregon, 2006.
Travel Award ($200) 2006 SAA Annual Meeting, San Juan, Puerto Rico,
Department of Anthropology, University of Oregon, 2006. Research Award ($500): Inferring Human Response to Late Holocene Ecological
Change in the Southern San Joaquin Valley, California, with Freshwater Mussel Shell Isotopes, Graduate School, University of Oregon, 2005.
Luther C. Cressman Prize for Outstanding Graduate Research Paper,
Characterization of Freshwater and Marine Radiocarbon Corrections at Elk Hills, Kern County, California, Department of Anthropology, University of Oregon, 2005.
viii
PEER-REVIEWED PUBLICATIONS:
Kennett, D.J., Culleton, B.J., 2012. A Bayesian Chronological Framework for Determining Site Seasonality and Contemporaneity. In Seasonality and Human Mobility along the Georgia Bight: Proceedings of the Fifth Caldwell Conference, St. Catherines Island, Georgia. E.J. Reitz, I.R. Quitmyer & D.H. Thomas (eds.), pp. 37-49. Anthropological Papers of The American Museum of Natural History No. 97.
Culleton, B.J., Prufer, K.M., Kennett, D.J., 2012. A Bayesian AMS 14C
Chronology of the Classic Maya Center of Uxbenká, Belize. Journal of Archaeological Science 39:1572-1586.
Rick, T.C., Henkes, G.A., Lowery, D.L., Colman, S.M., Culleton, B.J., 2012.
Marine Radiocarbon Reservoir Corrections (∆R) for Chesapeake Bay and the Middle Atlantic Coast of North America. Quaternary Research 77:205-210.
Prufer, K.M., Moyes, H., Culleton, B.J., Kindon, A., Kennett, D.J., 2011.
Formation of a Complex Polity on the Eastern Periphery of the Maya Lowlands. Latin American Antiquity 22:199-223.
Kennett, D.J., Culleton, B.J., Voorhies, B., Southon, J.R., 2011. Bayesian
Analysis of High-Precision AMS 14C Dates from a Prehistoric Mexican Shellmound. Radiocarbon 53:245–259.
Rick, T.C., Culleton, B.J., Smith, C.B., Johnson, J.R., Kennett, D.J., 2011. Stable
Isotope Analysis of Dog, Fox, and Human Diets at a Late Holocene Chumash Village (CA-SRI-2) on Santa Rosa Island, California. Journal of Archaeological Science. 38:1385-1393
Erlandson, J.M., Rick, T.C., Braje, T.J., Casperson, M., Culleton, B.J., Fulfrost,
B. Garcia, T., Guthrie, D., Jew, N., Kennett, D.J., Moss, M.L., Reeder, L., Skinner, C., Watts, J., Willis, L., 2011. Paleoindian Seafaring, Maritime Technologies, and Coastal Foraging on California’s Channel Islands. Science 331:1181-1185.
Jones, T.L., Culleton, B.J., Larson, S., Mellinger, S., Porcasi, J.F., 2011. Toward a
Prehistory of the Southern Sea Otter (Enhydra lutris nereis). In Human Impacts on Seals, Sea Lions, and Sea Otters: Integrating Archaeology and Ecology in the Northeast Pacific. T.J. Braje & T.C. Rick (eds.), pp. 243–271. University of California Press, Berkeley, CA.
McClure, S.B., García, O., Roca de Togores, C., Culleton, B.J., Kennett, D.J.,
2011. Osteological and Paleodietary Investigation of Burials from Cova de la Pastora, Alicante, Spain. Journal of Archaeological Science 38:420-428.
ix
Culleton, B.J., McClure, S.B., 2010. Análisis de la Paleodieta. In: Cavidades de Uso Funerario durante el Neolítico Final/Calcolítico en el Territorio Valenciano: Trabajos Arqueológicos en Avenc dels Dos Forats o Cova del Monedero (Carcaixent, Valencia). Archivo de Prehistoria Levantina Vol. XXVIII, pp. 193-194.
García Puchol, O., Cotino Vila, F., Miret Estruch, C., Pascual Benito, J. Ll.,
McClure, S.B., Molina Balaguer, Ll., Alapont, Ll., Carrión Marco, Y., Morales, J.V., Blasco Senabre, J., Culleton, B.J., 2010. Cavidades de Uso Funerario durante el Neolítico Final/Calcolítico en el Territorio Valenciano: Trabajos Arqueológicos en Avenc dels Dos Forats o Cova del Monedero (Carcaixent, Valencia). Archivo de Prehistoria Levantina Vol. XXVIII, pp. 139-206.
Kennett, D.J., Piperno, D.R., Jones, J.G., Neff, H., Voorhies, B., Walsh, M.K.,
Culleton, B.J., 2010. Pre-pottery Farmers on the Pacific Coast of Southern Mexico. Journal of Archaeological Science 37:3401-3411.
McClure, S.B., García Puchol, O., Culleton, B.J., 2010. AMS Dating of Human
Bone from Cova de la Pastora: New Evidence of Ritual Continuity in the Prehistory of Eastern Spain. Radiocarbon 52:25-32.
Lesure, R.G., Gagiu, A., Culleton, B.J., Kennett, D.J., 2010. Changing Patterns of
Shellfish Exploitation. In Settlement and Subsistence in Early Formative Soconusco: El Varal and the Problem of Inter-Site Assemblage Variation. R.G. Lesure (ed.), pp. 75-87. Cotsen Institute of Archaeology Press, Los Angeles.
Kennett, D.J., Culleton, B.J., 2010. Shellfish Harvesting Strategies at El Varal. In
Settlement and Subsistence in Early Formative Soconusco: El Varal and the Problem of Inter-Site Assemblage Variation. R.G. Lesure (ed.), pp. 173-178. Cotsen Institute of Archaeology Press, Los Angeles.
Kennett, D.J., Kennett, J.P., West, A., West, G.J., Bunch, T.E., Culleton, B.J.,
Erlandson, J.M., Que Hee, S.S., Johnson, J.R., Mercer, C., Shen, F., Sellers, M., Stafford, T.W., Jr., Stich, A., Weaver, J.C., Wittke, J.H., Wolbach, W.S., 2009. Shock-Synthesized Hexagonal Diamonds in Younger Dryas Boundary Sediments. Proceedings of the National Academy of Sciences USA 106(31):12623-12628.
Culleton, B.J., Kennett, D.J., Jones, T.L., 2009. Oxygen Isotope Seasonality in a
Temperate Estuarine Shell Midden: A Case Study from CA-ALA-17 on the San Francisco Bay, California. Journal of Archaeological Science 36:1354-1363.
x
Kennett, D.J., Kennett, J.P., West, G.J., Erlandson, J.M., Johnson, J.R., Hendy, I.L., West, A. Culleton, B.J., Jones, T.L., Stafford, T.W., Jr., 2008. Wildfire and Abrupt Ecosystem Disruption on California’s Northern Channel Islands at the Ållerød-Younger Dryas Boundary (13.0-12.9 ka). Quaternary Science Reviews 27:2528–2543.
Braje, T.J., Kennett, D.J., Erlandson, J.M., Culleton, B.J., 2007. Human Impacts
on Nearshore Shellfish Taxa: A 7,000 Year Record from Santa Rosa Island, California. American Antiquity 72:735-756.
Kennett, D.J., Culleton, B.J., Kennett, J.P., Erlandson, J.M., Cannariato, K.G.,
2007. Middle Holocene Climate Change and Population Dispersal in Western North America. In Climate Change and Cultural Dynamics: A Global Perspective on Mid-Holocene Transitions. D.G. Anderson, K.A. Maasch, & D.H. Sandweiss (eds.), pp. 531-557. Academic Press, San Diego, CA.
Culleton, B.J., Kennett, D.J., Ingram, B.L., Erlandson, J.M., Southon, J.R., 2006.
Intrashell Radiocarbon Variability in Marine Mollusks. Radiocarbon 48:387-400.
Culleton, B.J., 2006. Implications of a Freshwater Radiocarbon Reservoir
Correction for the Timing of Late Holocene Settlement of the Elk Hills, Kern County, California. Journal of Archaeological Science 33:1331-1339.
Newsome, S.D., Phillips, D.L., Culleton, B.J., Guilderson, T.P., Koch, P.L., 2004.
Dietary Reconstruction of an Early to Middle Holocene Human Population from the Central California Coast: Insights from Advanced Stable Isotope Mixing Models. Journal of Archaeological Science 31:1101-1115.
COMMENTS:
Culleton, B.J., 2008.Crude Demographic Proxy Reveals Nothing about Paleoindian Population (Comment on “Buchanan B. et al. (2008) Paleoindian Demography and the Extraterrestrial Impact Hypothesis. Proc Natl Acad Sci USA 105:11651–11654.”) Proceedings of the National Academy of Sciences USA 105:E111.
xi
ACKNOWLEDGMENTS In completing this dissertation, I am struck by what the turn-of-the-20th century
Russian philosopher-mystic George Ivanovich Gurdjieff called the process of ‘self-
remorse’ that comes with acknowledging the ‘conscious labors and intentional sufferings’
that our forebears took upon themselves to give us what we usually take for granted
today. He illustrates this emotion early in his vast allegory Beelzebub’s Tales to His
Grandson at the moment Beelzebub’s grandson Hassein first makes this realization:
“Only now have I come to understand that everything we have at the present time and everything we use – in a word, all the contemporary amenities and everything necessary for our comfort and welfare – have not always existed and did not make their appearance easily.
It seems that certain beings in the past have during very long periods labored and suffered very much for this, and endured a great deal which perhaps they even need not have endured.
They labored and suffered only in order that we might now have this and use it for our welfare.
And this they did, either consciously or unconsciously, just for us, that is to say, for beings quite unknown and entirely indifferent to them.
And now not only do we not thank them, but we do not even know a thing about them, but take it all as in the natural order, and neither ponder nor trouble ourselves about this question at all. ....
And so, my dear and kind Grandfather [Beelzebub] ... I have gradually, with all my presence, become aware of all this, the need to make clear to my Reason why I personally have all the comforts which I now use, and what obligations I am under for them.”
G.I. Gurdjieff (1973:76-77; emphasis added) I’ll attempt to make clear to whom I owe my comforts and to whom I am obliged
for them. First, thanks to my committee members Drs. Doug Kennett, Jon Erlandson,
Madonna Moss, Keith Prufer, and Pat Bartlein for constantly pushing me outside of my
intellectual comfort zones and into the hard work of doing original research. I am grateful
for your encouragement, friendship and inspiration. I am grateful to the People of Belize
for allowing me to conduct my dissertation research and sharing their cultural heritage
xii
with me. To my friends the Bardalez family in Big Falls for making me feel at home
away from home. To Mr. Don Owen-Lewis for putting up with me long enough to
develop an amicable rapport. Most of all I owe a deep debt of gratitude to the people of
Santa Cruz village who worked so hard with me digging all those trenches and traipsing
around weighing corn in typical loco gringo style. Your friendship, sense of humor, and
gracious hospitality under difficult circumstances is truly humbling. To the rest of the so-
called Golden Four: Josh Fisher, Emily Henderson-Guthrie, and Aaron Blackwell. To my
dear friends outside the department who kept me from being all-Anthro-all-the-time:
Colleen Laird, Carter Soles, Emily Afanador, Kom Kunyosying and Kelly Sultzbach.
Thanks to the Anthro Department staff that have always been there for
administrative matters: Tiffany Brannon, Steph Morton, Betina Lynn, and Leah Frazier.
And I salute Brenda Dutton and her unflagging Commitment to Functionality. To my
collaborators and co-authors on my dissertation work: Claire Ebert, Ethan Kalosky, Doug
Kennett, Keith Prufer, and Bruce Winterhalder. Claire crafted many of the best maps and
GIS coverages that appear in this dissertation. Thanks to the National Science Foundation
for supporting me with a Graduate Research Fellowship (GRFP-20060227) and a
Doctoral Dissertation Improvement Grant (DDIG-0829218), as well as a Research
Assistantship through Dr. Kennett’s Human Social Dynamics grant (HSD-0827305).
Most of all I thank my family for their unconditional support through this process.
To my Uncle Pat for being a solid friend over the last several years. To my late father
Patrick for instilling me with curiosity for archaeology and the natural sciences. To my
mother Kitty, who worked so hard to give me and my sister the best opportunities we
could have.
xiii
This work is dedicated to: My mother, Catherine P. Culleton, who supported me unconditionally during my
research;
My father, Patrick G. Culleton, whose interest in nature and archaeology inspired me to pursue this line of work; and
My great uncle, Monsignor James H. Culleton, a great historian and man of the church
who inspired my father in his youth.
xiv
TABLE OF CONTENTS
Chapter Page I. INTRODUCTION .................................................................................................... 1
Setting and Background of Uxbenká ..................................................................... 5
Dissertation Fieldwork ........................................................................................... 11
Organization of the Dissertation ............................................................................ 12
II. A BAYESIAN AMS 14C CHRONOLOGY OF THE CLASSIC MAYA
CENTER OF UXBENKÁ, BELIZE...................................................................... 16
The Setting of Uxbenká in the Maya Lowlands .................................................... 18
Methods .............................................................................................................. 24
Radiocarbon Sampling and Measurement ....................................................... 24
The Bayesian Framework ................................................................................ 26
Results .................................................................................................................... 29
Group A (Stela Plaza) ...................................................................................... 29
Group B ............................................................................................................ 36
Group D ........................................................................................................... 40
Discussion .............................................................................................................. 44
Conclusions ............................................................................................................ 49
III. CHANGING AGRICULTURAL AND URBAN LANDSCAPES AT THE
CLASSIC MAYA CENTER OF UXBENKÁ, BELIZE ....................................... 51
xv
Chapter Page Climatic Context .................................................................................................... 55
The History of Maya Land Use ............................................................................. 57
Field Methods ........................................................................................................ 61
Chronology ............................................................................................................ 61
Results .................................................................................................................... 63
Bedrock Geology and the Geoarchaeology of Uxbenká .................................. 63
Excavations in the Site Core ............................................................................ 65
Excavations Northeast of the Site Core, Cochil Bul ........................................ 74
Discussion .............................................................................................................. 77
Conclusions ............................................................................................................ 85
IV. MAIZE AGROECOLOGY AND POPULATION ESTIMATES FOR THE
ANCIENT MAYA POLITY OF UXBENKÁ, BELIZE ....................................... 87
Setting and Background ......................................................................................... 93
Methods.................................................................................................................. 98
Maize Yields and Modern Populations .................................................................. 105
Population Estimates for the Uxbenká Polity ........................................................ 110
Discussion .............................................................................................................. 112
Conclusions ............................................................................................................ 118
xvi
Chapter Page V. THE IDEAL FREE AND DESPOTIC DISTRIBUTIONS AND ANCIENT
MAYA SETTLEMENT AT UXBENKÁ, BELIZE .............................................. 120
The Ideal Free and Despotic Distributions ............................................................ 122
Archaeological Applications of the IFD and IDD ................................................. 127
Applying the IFD to Household Settlement at Uxbenká ....................................... 131
Agricultural Productivity ................................................................................. 133
Hydrology ........................................................................................................ 135
Distance from the Site Core ............................................................................. 136
The Model .............................................................................................................. 138
Further Work towards Understanding the Ideal Free and Despotic Settlement
Models for Uxbenká ........................................................................................ 146
Conclusions ............................................................................................................ 148
VI. CONCLUSIONS AND PROSPECTS ................................................................... 150
Broader Relevance to Lowland Maya Archaeology .............................................. 156
The Archaeology of Uxbenká and the Community of Santa Cruz ........................ 162
REFERENCES CITED ................................................................................................ 165
xvii
LIST OF FIGURES Figure Page 1.1. Location of Uxbenká in relation to Lowland Maya sites ..................................... 6 1.2. Location of Uxbenká and other Maya sites in southern Belize ........................... 7 2.1. Detail maps showing excavations at: A) the Uxbenká site core; B) Group A (Stela Plaza); C) Group B; and D) Group D. ....................................................... 27
2.2. Profile of Unit 1, SubOp 08-4 on Str. A1 ............................................................ 32
2.3. Profile of SubOp 07-5 on Str. A1. ....................................................................... 33
2.4. Profile of SubOp 07-3 on Str. A6 ........................................................................ 35
2.5. Profile of Unit 2, SubOp 08-7 in Group B ........................................................... 38
2.6. Profile of Unit 1, SubOp 09-14 in the Group D plaza ......................................... 42
2.7. Summary of modeled calibrations for key construction episodes at Groups A, B and D............................................................................................... 45
3.1. The Uxbenká site core, showing locations of geoarchaeological excavations in A) the core and B) the Cochil Bul area to the north. ....................................... 54
3.2. Characteristic exposures of the Toledo Beds in the Uxbenká vicinity ................ 64
3.3. Transects showing the marked jointing in the bedrock. ...................................... 66
3.4. Composite profile of AG3 and AG4 west walls, showing sandstone alignment (Feature 1). .......................................................................................... 67
3.5. AG8 west wall profile, showing sandstone alignment (Feature 1) ...................... 70
3.6. AG9 west wall profile. ......................................................................................... 73
3.7. Composite profiles of Cochil Bul excavations AG12, AG13, and Auger Probe ......................................................................................................... 76
xviii
Figure Page 3.8. Geomorphic stability and instability at Uxbenká compared to cultural chronology and climate records ........................................................................... 78
4.1. Scatterplots of edible and waste maize vs. whole ear weight for test plots ......... 100
4.2. Relationship between bulk maize yield (kg/ha) and number of plantings per sample plot. .................................................................................................... 101
4.3. Comparison of 2009 and 2010 density-normalized maize yields ........................ 103
4.4. Scatterplots of bulk maize yields vs. environmental variables ............................ 104
4.5. Scatterplots of normalized maize yields vs. environmental variables ................. 104
4.6 Interpolated raster of maize yields around Santa Cruz village ............................ 107
4.7. Hypothetical catchments centered on Uxbenká used to estimate the total maize production for each 1km radius. ................................................................ 111
5.1. Habitat rankings under assumptions of the Ideal Free Distribution and with the Allee effect ............................................................................................. 124
5.2. Composite showing ecological variables incorporated into the IFD model of settlement at Uxbenká ..................................................................................... 134
5.3. Examples of high and low ranked settlement groups based on the ecological variables within a 0.5 km catchment radius ....................................... 140
xix
LIST OF TABLES Table Page 2.1. Lowland Maya chronological periods ............................................................... 18
2.2. AMS 14C dates from Uxbenká used in Bayesian modeling ............................... 25
2.3. Examples of OxCal commands and relevant stratigraphic situations ................ 29
2.4. Stela dates from Group A, Uxbenká .................................................................. 30
2.5. Modeled results for three Group A stratigraphic sequences .............................. 36
2.6. Modeled results for the Group B stratigraphic sequence ................................... 39
2.7. Modeled results for the Group D stratigraphic sequence ................................... 43
3.1. Calibrated AMS 14C dates from paleosols ......................................................... 62
3.2. Geochronology of paleosols in the Uxbenká site core....................................... 77
4.1. 2009 bulk, edible and normalized maize yields ................................................. 101
4.2. 2010 bulk, edible and normalized maize yields ................................................. 102
4.3. Regression of 2009 maize yield normalized for planting density with respect to environmental variables..................................................................... 105
4.4. Regression of 2010 maize yield normalized for planting density with respect to environmental variables..................................................................... 105
4.5. Regression of 2-year aggregated maize yield data normalized for planting density with respect to environmental variables. ................................. 105
4.6. Estimated annual maize yield and potential population of Santa Cruz Village ................................................................................................................ 110
4.7. Estimated potential population of Santa Cruz Village assuming % of land devoted to maize ........................................................................................ 110
4.8. Estimated maximum population for Uxbenká assuming different catchment areas. ................................................................................................. 112
xx
Table Page 4.9. Estimated population of Uxbenká at various catchments and fallow length assuming 75% maize cultivation ............................................................ 113
4.10. Population estimates for Lowland Maya sites ................................................... 116
5.1. Chronological data on 22 settlement and core groups considered in IFD modeling ....................................................................................................... 132
5.2. Environmental parameters used to rank settlement groups and core groups at Uxbenká ............................................................................................... 139
5.3. Chronological data and site rankings by individual variables and overall rankings .................................................................................................... 141
1
CHAPTER I
INTRODUCTION
Characterizing the tempo and mode of ancient landscape transformation under
intensive agricultural production in relation to social, ecological and climatic change is
crucial to understanding the development and disintegration of complex societies
(deMenocal 2001; Redman 1999, 2005; van der Leeuw and Redman 2002).
Intensification, defined here as increased labor input into resource acquisition or
production, is a crucial component in the emergence of complex societies, which share
the hallmarks of social hierarchy, differential access to resources, division of labor,
technological elaboration, and craft specialization (Boserup 1965; Carneiro 1970; Earle
1987; Flannery 1972; Friedel and Schele 1988; Netting 1993; Price and Gebauer 1995).
Intensification increases productive capacity of a landscape by increasing production
yields per unit area while decreasing returns on labor. This can provide an economic
foundation for population growth, relatively stable and persistent forms of land tenure,
and social resilience (Redman 2005). Contemporary societies, while comprising multiple
scales of both extensive and intensive resource production and distribution, are
increasingly viewed as subject to the same external and internal disturbances that
transformed ancient societies (e.g., prolonged drought, anthropogenic environmental
degradation, demographic crisis, warfare, disease). The growing public concern over the
fate of societies places great demands on archaeologists and the archaeological record to
go beyond descriptions of ‘collapses’ to explain the processes by which complex
societies emerge, persist, develop and disintegrate in the context of the changing human
2
ecosystem and provide insights into our future prospects (Redman 1999, 2005; van der
Leeuw and Redman 2002).
In the last two decades, natural and anthropogenic environmental change has
gained predominant explanatory weight in the disintegration of Classic Lowland Maya
polities. Stretching from the Petén (Guatemala) through Belize and into Mexico’s
Yucatán peninsula, the Maya Lowlands saw the emergence of socially stratified and
politically complex societies from the Middle Preclassic (1000/800 – 400 BC) through
the Terminal Classic (AD 80 – 1000) periods (Demarest 2004:8-12). Maize farming
formed the main economic and cultural basis of Maya society, supplemented by cultivars
such as beans, manioc, ramón nuts, and cacao, as well as a variety of hunted and gathered
wild game, fish, and plant resources. Transformation of the subtropical and tropical
lowland forests through human use has a long history that continues to be elucidated.
High resolution regional climate and geomorphic records have provided evidence for
deforestation and erosion associated with urbanization, and extended droughts throughout
the Classic Period that undermined already fragile productive capabilities, contributing to
the Classic Maya ‘collapse’ (Culbert 2004; Curtis et al. 1996; Deevey et al. 1979;
Demarest 2006; Gill 2000; Haug et al. 2001, 2003; Hodell et al. 1995, 2001; Webster
2002). In danger of getting lost in the hype of Maya megadroughts is the fact that
multiple land use strategies, conservative and otherwise, were employed throughout the
Maya region that allowed for the development and stability of polities for extended
periods. In the part of the Maya lowlands where water sources were localized in cenotes
(e.g., the Yucatán), or highly alkaline (e.g., the Petén), drought was likely a dominant
factor in the Classic Period “collapse” (e.g., Lucero 2002; Scarborough 2003), perhaps so
3
dominant as to obscure the other ecological and social dynamics leading to Terminal
Classic political disintegration. Tracking Maya-environment interactions in the wetter
climate and richer soils of southern Belize may de-emphasize the role of adverse effects
of natural and human-induced environmental change, allowing the complexities of land
use decisions, intensification strategies and demographic change to be more clearly
understood.
Through the study of site-specific adaptations to local environmental conditions,
climate regimes, and social development, I seek to understand how individual land use
decisions permitted growth of these polities and elaboration of social forms within a
changing ecological context. This approach reflects my belief that while archaeology
(and anthropology) has the ability to address problems relevant to human societies at any
spatial or temporal scale, the observations used to develop and test theories must be
empirically grounded and oriented at human scales of perception and action to be
informative. Individual subsistence farmers, for example, may take into account
environmental and sociopolitical conditions at regional scales over the long term, but
deciding when to clear fields, when and what to plant, how to allocate labor, and so on,
are often dictated by local conditions and immediate-term considerations.
Focusing on individual decision-making in terms of land use and settlement in the
archaeological record is achieved in parts of this study through the use of models from
Human Behavioral Ecology (HBE), which, as a part of Evolutionary Ecology, maintains
methodological individualism as a central concept (Smith and Winterhalder 1992:39-41;
Winterhalder 1994; Winterhalder and Smith 1992). As described by Smith and
Wintehalder (1992:39), “[m]ethodological individualism ... holds that properties of
4
groups (social institutions, populations, societies, economies, etc.) are a result of the
actions of its individual members”, and that explanations of group actions should
necessarily be built “from the bottom-up”. At Uxbenká and most Classic Maya centers,
this places the individual commoner and their household unit as the focus of most land
use decision making, while acknowledging the potential for top-down management of
group resources (land, labor, agricultural production, etc.) by elites. While some aspects
of HBE have been argued to be overly reductive (Winterhalder and Smith 1992:23),
methodological individualism can serve as a theoretical bridge between processual and
post-processual approaches, specifically in regards to critiques of the former for failing to
acknowledge individual agency in negotiating social formations including gender and
class (e.g., Brumfiel 1992).
HBE provides a coherent framework in which to integrate a diverse array of
social, ecological and historical data to build models of past behavior and to generate
testable hypotheses about the archaeological record. They also are flexible and readily
generalizable, so that insights gained at smaller spatial and temporal scales (or with more
simplified models) can be applied and tested at larger scales (or with more complex
models) (Winterhalder and Kennett 2006).
To build a context for the application of HBE models at the ancient Maya center
of Uxbenká, this study synthesizes data from the broad archaeological literature on the
Lowland Maya, and presents new data and analysis of: architectural chronology
beginning in the Late Preclassic Period (60 BC – AD 220) through the Late Classic
Period (AD 600-800); the geoarchaeological record of changing land use in the site core
from the early Middle Preclassic Period (ca. 970 BC – 620 BC) through the Terminal
5
Classic Period (AD 800-900); empirical data on contemporary maize yields surrounding
Uxbenká and the nearby Maya village of Santa Cruz to develop estimates of past
population density at Uxbenká; and development and testing of an ecologically-based
predictive model of settlement through the center’s history. To orient the reader with the
region where Uxbenká is found, I provide the following background summary, which is
elaborated upon in each of the remaining chapters
Setting and Background of Uxbenká
Southern Belize is home to diverse geologic and ecological zones, from the Maya
Mountains to the west, into the foothills that host the primary ancient Maya centers of
Pusilhá, Uxbenká, Lubaantun, and Nim Li Punit (Figures 1.1, 1.2), across the narrow
strip of coastal plains to the mangrove swamps and lagoons of the coast where Maya salt
production and maritime trade flourished during the Classic and Postclassic periods
(McKillop 2008). The Maya Mountains served as a natural boundary separating southern
Belize from the rest of the ancient Maya world and are composed of a mixture of
Cretaceous intrusive and extrusive rocks, volcanics (e.g. rhyolite and welded tuffs) and
metavolcanics (Bateson and Hall 1977). People living in the immediate vicinity of these
durable rocks (e.g., Ek Xux and Muklebal Tzul; Abramiuk and Meurer 2006) capitalized
on these raw materials for the manufacture and trade of groundstone milling tools from
the Late Preclassic through Terminal Classic periods (50 BC-AD 1000; Prufer et al.,
2011).
6
Figure 1.1. Location of Uxbenká in relation to Lowland Maya sites discussed in the text (map by C. Ebert).
7
Figure 1.2. Location of Uxbenká and other Lowland Maya sites in southern Belize (map by C. Ebert).
The foothills of the Maya Mountains are comprised of Cretaceous limestones and
a series of interbedded Tertiary marine sediments known locally as the Toledo Beds
(synonymous with the Sepur Formation in Guatemala) (Keller et al. 2003). The Toledo
Beds range from shallow water limestones and dolomites to deeper water calcareous
8
shales, mudstones and sandstone members (Keller et al. 2003; Miller 1996; Wright et al.
1959).
Most of the known ancient Maya centers in the region are set on these Tertiary
sediments, with the exception of Pusilhá in the south, which is set on a Cretaceous
limestone. Uxbenká and Pusilhá were the earliest centers established during the Late
Preclassic (AD 20-200; Culleton et al. 2012; Prufer et al. 2011) and Early Classic (AD
300-600; Braswell et al. 2004), respectively. The chronologies of Lubantuun and Nim Li
Punit are less well-understood, but both appear to be restricted to the Late Classic (AD
600-800; Hammond 1975; Hammond et al. 1999).
The character of the Toledo Beds is expressed differently at each of these
locations – massive sandstone beds form natural stone plazas at Nim Li Punit, for
example, but extensive mudstone and sandstone outcrops form natural terraces
surrounding Uxbenká. The zone around Uxbenká is close to a discontinuity between
Cretaceous and Tertiary members of the Toledo Formation, expressed most notably by a
prominent Cretaceous limestone karst ridge immediately south of the site and the present
village of Santa Cruz (Keller et al. 2003; Miller 1996). The karstic ridge, locally known
as “The Rock Patch”, contains several caves that are the subject of on-going
archaeological research (Prufer et al. 2011). This karst ridge dominates the drainage of
the Rio Blanco, which flows with its tributaries over the Tertiary beds south until meeting
the southwest-northeast trending ridge where it abruptly turns to the east. Eventually the
Rio Blanco enters the karst at Oke’bal Ha Cave and exits as Blue Creek to the south at
Hokeb Ha Cave (Miller 1996).
9
The coastal plain between the foothills and Caribbean Sea is made up primarily of
Pleistocene fluvial sediments discharged from drainages originating in the Maya
Mountains and associated foothills. North of Deep River the soils are rocky, heavily
weathered and covered in open pine savannah. Aside from a few sites along rivers
(Graham 1994), ancient Maya settlements on the savannah are unknown, presumably
because the soils are poorly suited to maize agriculture. South of the Monkey River the
pine savannah gives way to cohune palm (Attalea cohune) forest where many of the
modern villages are located and are founded upon maize, citrus, rice and ground-crop
cultivation. The earliest human occupations in this zone are poorly understood, but a
fluted point found near the village of Big Falls on the Rio Grande suggests Paleoindian
(~13,000 cal. BP) activity (Lohse et al. 2006; Weintraub 1994). Several Archaic Period
projectile points (Lowe Points; 2500-1900 BC) have also been found in plowed fields
between Big Falls and the village of Hiccatee to the north, but prehistoric settlement is
generally limited to scattered evidence of small Classic period settlements.
Mangrove swamps and brackish lagoons that are largely inaccessible without
watercraft characterize the coastal zone of southern Belize. Sea level rise and
stabilization is implicated in the formation of extensive mangrove swamps during the
middle Holocene, from 6000-3000 cal BC (McIntyre et al. 2004; Wooller et al. 2007).
The earliest documented settlements date to the Early Classic (AD 300-600). These
include the modest sites of Butterfly Wing (mouth of Deep River) and Wild Cane Cay
that both suggest the exploitation of marine and estuarine resources and early maritime
activities (Graham 1994; McKillop 1996, 2010). The Late Classic Period saw the rise of a
salt production industry and maritime trade that persisted into the Postclassic (McKillop
10
1995, 1996, 2005a, 2005b, 2008, 2010). Remains of wooden salt processing stations,
structures, weirs and a canoe paddle have been documented at several coastal sites that
have been submerged and preserved by ~1m of relative sea level rise since the Late
Classic. Obsidian artifacts from Guatemalan sources found at Wild Cane Cay also
suggest that these coastal sites were engaged with overland trade networks likely
facilitated by up-river canoe travel to the west as well as maritime connections with the
north to the Belize River Valley and coastal Yucatán (McKillop 1989).
Monsoonal rains largely drive erosion and deposition on the coastal plain.
Precipitation decreases from the coast (~4000 mm/yr) to the interior (2400 mm/yr) as
elevation increases (Heyman and Kjerfve 1999). Annual rainfall in the area of Santa Cruz
village, the location of Uxbenká, is estimated at ~2700-3400 mm/yr. The annual climate
cycle of southern Belize is marked by distinct wet and dry seasons with relatively little
seasonal variation in temperature through the year. The onset of wet conditions differs
from year to year, but typically runs from June through September, when monthly rainfall
ranges from 400-700 mm (Hartshorn et al. 1984; Heyman and Kjerfve 1999; Wright et al.
1959). A short (2-3 weeks) dry spell known as the “canicula” often occurs in August
(Wright et al. 1959). The months of February through April are the driest months
(averaging 40-70mm/mo), and this is the period in the traditional milpa cycle when forest
is cleared for the wet season crops of maize, beans and other “ground foods” (Heyman
and Kjerfve 1999; Wright et al. 1959). The hurricane season, as elsewhere in the tropics,
occurs between August and October. Southern Belize is largely shielded from easterly
winds by the highlands of northern Honduras so hurricanes rarely make landfall.
Hurricane Iris in October 2002 was a devastating exception that left roughly 10,000
11
people homeless in Toledo District and destroyed that year’s wet season milpa crops
(Zarger 2002:xii-xiii).
Dissertation Fieldwork
Fieldwork for this dissertation was carried out at Uxbenká and the lands around
Santa Cruz village over multiple field seasons from June 2006 to October 2010. The
initial reconnaisance trip to Uxbenká in June 2006 was focused on identifying and
recovering speleothems from Yok Balum Cave with Doug Kennett, Kevin Cannariato
and Keith Prufer. I observed the possible terrace features in the Uxbenká at this time, and
planned for a preliminary season of excavations in 2007. The 2007 season was broken up
into a 3-week trip from February to March recovering sediment cores from around
Toledo District, and a 5-week trip from May to June to conduct preliminary excavations
on the presumptive terraces in the site core.
The 2008 field effort spanned 9 weeks from April to June where I made extensive
geoarchaeological excavations in the site core area, and did geologic reconnaissance of
the areas to the east of the the site core (i.e., settlement groups 25-28). The 2009 field
season ran for 10 weeks from April to June, being split between additional
geoarchaeological excavations in the site core and Cochil Bul area, water retention
features to the east of the site core, and setting maize yield plots in the milpas around
Santa Cruz. I returned for 3 weeks in September and October of the year to quantify the
yields in those plots. I carried out settlement excavations at five settlement groups from
April to June 2010, as well as establishing that seasons maize yield plots, which were
revisited in October for 3 weeks to collect the final yield data.
12
Organization of the Dissertation
Chapter II describes integration of high-resolution AMS radiocarbon dates with
stratigraphic information from selected archaeological sequences into a Bayesian
framework to produce a new chronology of major construction events in the Uxbenká site
core. The data are drawn from four seasons of investigations by members of the Uxbenká
Archaeological Project directed by Dr. Keith M. Prufer (University of New Mexico)
using contextual information produced by his team and radiocarbon samples processed at
the University of Oregon Archaeometry Facility and measured at The Keck Carbon Cycle
AMS Facility at UC Irvine. The calibration program OxCal was used to devise Bayesian
models that allowed for events that are not directly dated – such as the initial clearing of a
building site or the placement of a plaster floor – to be estimated based on its
stratigraphic relationship to directly dated events in well-constrained sequences. It can
also trim the calibrated ranges of directly dated events, which is advantageous within the
reversals in the radiocarbon curve during the Classic Period (AD 300 – 900). Results of
this analysis indicate earlier initial construction of three main architectural groups in the
site core than previously supposed, with the main plaza established during the Late
Preclassic Period (60 cal BC – cal AD 220), and continued remodeling and replastering
of the groups into the Early Classic Period (cal AD 300-600). These results confirm
Uxbenká as the earliest known Maya center in southern Belize (Prufer et al. 2011), and
point to specific areas of the Late Classic Period chronology to be refined by further work
at the site. This chapter, prepared as a co-authored work with Dr. Prufer and Dr. Douglas
J. Kennett, is published in the Journal of Archaeological Science.
13
Chapter III presents the results of geoarchaeological work within the Uxbenká site
core from the early Middle Preclassic Period (ca. 970 BC – 620 BC) through the
Terminal Classic Period (AD 800-900). Paleosols indicate human activity, land
clearance, and erosion consistent with swidden agriculture starting in the Middle
Preclassic Period, and provide the earliest evidence of ceramics in southern Belize. The
urban landscape during the Early and Late Classic periods (AD 300-800) was notably
stable, possibly due to the relocation of milpas outside the city center. The absence of
terraces in this hilly landscape suggests that swidden cultivation remained viable without
these labor investments throughout the Classic Period. Increased erosion and landscape
instability in the urban core during the Terminal Classic Period (AD 800-900) suggests
that the area was largely abandoned in terms of permanent settlement by that time, and
the land had reverted to swidden cultivation by a remnant farming population. This
chapter was prepared as a co-authored work with Dr. Prufer and Dr. Kennett, and has
been submitted to Geoarchaeology: An Interdisciplinary Journal.
Chapter IV describes the quantification of maize yields under swidden cultivation
by the contemporary Maya farmers of Santa Cruz village, on whose lands the ruins of
Uxbenká are located. Yield data were collected in 2009 and 2010 and compared to soil
nutrient and landscape characteristics to identify areas of greatest desirability for
household settlements around Uxbenká during its florescence. These data were then used
to estimate the potential maximum population that could be sustained under different
scenarios of overall productivity, fallow length, and level of intensification. Maximum
population of Uxbenká during the Classic Period was estimated to range between 7500
and 13,000 people within the 6km radius that could have been under its political
14
influence. This population is modeled at a five-year fallow period, on the verge of what
would be a true short fallow system, and suggests a low level of agricultural
intensification consistent with the lack of terracing and other similar features described in
Chapter III. The estimate of household settlement density predicted here can be tested
against future work in household, settlement and landscape archaeology at Uxbenká. This
chapter was prepared as a co-authored work with Dr. Bruce Winterhalder (UC Davis),
Claire Ebert (Penn State University), Dr. Prufer, and Dr. Kennett, and will be submitted
to Human Ecology.
Chapter V presents the development and testing of a population ecology model of
settlement expansion around Uxbenká based on the Ideal Free Distribution (IFD) and
related Ideal Despotic Distribution (IDD). The locations of 22 known civic/ceremonial
architectural groups and household settlement groups were ranked based on three
measures of suitability: agricultural potential (using the maize yield data presented in
Chapter IV), access to fresh water, and proximity to the site core. The prediction of the
IFD model is that the highest ranked habitats should be settled first, and as population
density increases, settlements will expand into less favorable habitats over time.
Comparison of the existing archaeological chronology with settlement ranks shows a
general conformity with the IFD, in that several of the earliest Late Preclassic settlements
are found in high-ranked locations near the site core, and in the most agriculturally
productive areas away from the site core. The location of a substantial and early civic-
ceremonial group (Group I) confounds the predicted pattern and is found in a much
lower-ranked habitat to the west of the Uxbenká’s urban core than is predicted by the
model. The presence of Group I in a marginal habitat early in the settlement history of
15
Uxbenká may be indicative of hierarchical conditions best described by the IDD and
suggests competitive exclusion of the site core by a ruling elite. This suggests that status
rivalry between competing elites played a significant role in the social geography and
settlement history of the site as early as the Late Preclassic. The results of this analysis
demonstrate the utility of formal IFD and IDD models to define ecological and social
factors affecting population distributions in the ancient Maya Lowlands and to identify
and explain instances of status competition more broadly in the archaeological record.
This chapter was prepared as a co-authored work with Dr. Winterhalder, Ms. Ebert,
Ethan Kalosky (University of New Mexico), Dr. Prufer, and Dr. Kennett, and will be
submitted to Journal of Anthropological Archaeology.
Chapter VI summarizes the major findings of this dissertation and places them
within a broader methodological and theoretical context.
16
CHAPTER II
A BAYESIAN AMS 14C CHRONOLOGY OF THE CLASSIC MAYA CENTER OF
UXBENKÁ, BELIZE
This work was published in volume 39 of the Journal of Archaeological Science
in May 2012. Keith M. Prufer provided access to stratigraphic information and
excavation profiles from the archaeological work at Uxbenká. I processed the
radiocarbon samples, evaluated the existing chronometric database, and incorporated the
chronological and stratigraphic information in a Bayesian framework for analysis and
interpretation. Douglas J. Kennett provided guidance and original insights into the
crhonological interpretations. I was the principle investigator for this work.
Archaeological research in the Maya region is heavily dependent upon ceramic
typologies to estimate the age of sites. In parts of the Maya lowlands where these
typologies are well-established (e.g., central Petén, Belize Valley) they are used to
determine relative cultural sequences and sometimes rough estimates of absolute age
(e.g., Culbert and Rice 1990; Demarest et al. 2004). Ultimately these age estimates are
based on older uncalibrated 14C dates and the large error margins of these older 14C dates
make some of the finer grained ceramic age estimates (sometimes shorter than 50 years)
unrealistic. In the last two decades archaeologists have employed multiple
complementary (or alternative) chronometric techniques to augment and refine ceramic-
based chronologies in Mesoamerica, including archaeomagnetism (Wolfman 1990);
obsidian hydration (Webster et al. 2004); epigraphy (LeCount et al. 2002), and AMS 14C
dating, to improve site chronologies and the age estimates of certain ceramic types
17
(Garber et al. 2004; Healy 2006; LeCount et al 2002; Moyes et al. 2009; Prufer et al.
2011; Rosenswig and Kennett 2008; Saturno et al. 2006; Webster et al. 2004). Several
major analytical and statistical improvements in AMS 14C dating and calibration now
allow more precise chronological estimates that sometimes approach +/- 15-30 calibrated
years under ideal circumstances (Kennett et al. 2011). Precise and accurate age
determinations are necessary to compare cultural sequences against high-resolution
historical, environmental and climatic datasets as archaeologists in the Maya region ask
increasingly sophisticated and relevant historical, demographic and environmental
questions (e.g., Beach et al. 2009, Braswell 2003; Demarest et al. 2004; Aimers and
Hodell 2011; Turner 2010; Lentz and Hockaday 2009; Webster 2002). Identifying causal
relationships between social and environmental effects in these records requires
chronological precision capable of establishing the true order of those events, and ideally
discerning whether events are actually contemporaneous (Marcus 2003:344-345).
In this chapter I build upon the growing number of AMS 14C studies in the Maya
region and the work of the Uxbenká Archaeological Project (Prufer et al. 2011) by
employing a Bayesian chronological framework to generate a more precise chronology
for the growth and contraction of this Classic Maya polity in southern Belize. The
Bayesian analysis of radiocarbon dates from archaeological sites is becoming routine in
Britain (Buck 2004; Bayliss and Bronk Ramsey 2004; Bayliss et al. 2007) and programs
like OxCal (Bronk Ramsey 1995, 2001, 2005, 2009) provide a prepackaged set of
Bayesian statistical tools to help develop finer-grained archaeological site chronologies.
Having been the focus of an intensive high-precision radiocarbon dating program for
several years (Prufer et al. 2011), the site provides a unique opportunity to apply a
18
Bayesian approach to a Lowland Maya site, and demonstrate the potential for broader
applications in the Maya region and elsewhere in Mesoamerica. First I review of the
regional archaeological chronology, and then give a basic overview of the Bayesian
approach to incorporating archaeological observations with radiocarbon data using
OxCal. These techniques are applied to a sample of the Uxbenká AMS radiocarbon
database to investigate the tempo of development and decline at the site based on the
available data.
The Setting of Uxbenká in the Maya Lowlands
While noting that regional chronologies differ in the timing of Lowland Maya
culture-historical phases, the temporal units discussed in this chapter generally follow
Demarest’s (2004:13) chronological scheme (Table 2.1). Because of the relatively late
development of ceramic technology in Belize, the Late Archaic is considered to extend
until ca. 1000-800 BC in local or sub-regional contexts (see Lohse 2010).
Table 2.1. Lowland Maya chronological periods (after Demarest 2004:13 and Lohse 2010)
Period Span Late Archaic 3000 BC – 1000-800 BC Middle Preclassic 1000-800 BC - 400 BC Late Preclassic 400 BC - AD 300 Early Classic AD 300 - AD 600 Late Classic AD 600 - AD 800 Terminal Classic AD 800 - AD 1000 Early Postclassic AD 1000 - AD 1300 Late Postclassic AD 1300 - AD 1519
When Uxbenká was first settled it was positioned in a geopolitically marginal
region. Through time it found itself situated near trade routes connecting larger polities,
including Tikal, Copán, and Caracol (see Figure 1.1). The temporal span considered in
19
this chapter covers the latter portion of the Late Preclassic Period (ca. 100 BC - AD 300),
through the Classic Period (AD 300-1000). The Late Preclassic witnessed both the
development and disintegration of major political centers in the central Maya Lowlands,
with massive expansion and political centralization occurring at Tikal and Calakmul
corresponding with a decline of authority at the earlier power centers of Nakbe and El
Mirador (Folan et al. 1995; Hansen 2006; Harrison 2006; also Martin and Grube 2008).
The Early Classic in the Petén is characterized by the ascendancy of Tikal as a regional
power, and the extension of its influence southward towards Copán around AD 426
(Sharer 2003: 322). Tikal's greater regional influence was possibly stimulated by
increased interaction after AD 378 with the highly centralized and expansionistic state of
Teotihuacan located in the central Mexican highlands (see Braswell 2003). This would
have facilitated Tikal’s access to lucrative trade routes in the southern Petén and
southeastern lowlands (Sharer 2003: 351).
Southern Belize is located in a geographic and cultural frontier of the Maya
Lowlands. Like other Maya frontiers (Henderson 1992), it was both peripheral to, yet
connected with the cultural and political developments occurring in larger and more
economically and politically powerful centers (Schortman and Urban 1994). During
southern Belize’s apogee between AD 400-900 its polities were involved in a variety of
trade and exchange activities, focused on mineral and biotic resources (e.g., groundstone,
cacao, clays for ceramics production; Abramiuk and Meurer 2006; Dunham 1996;
Dunham and Prufer 1998; Graham 1987), agricultural production (Prufer 2005a), and
marine resources that linked polities from the Petén to the Caribbean Sea (Hammond
1978; McKillop 2005a, 2005b).
20
Until recently most regional settlement chronologies relied on architectural
features (e.g., ballcourts), epigraphic data, and to a lesser extent comparison of ceramics
with other regions of the Maya Lowlands (e.g. Dunham 1996; Hammond 1975;
Leventhal 1990, 1992). In general, these studies indicate that the number of polities and
density of settlements were highest during the Late Classic. To the north of Uxbenká, in
the Stann Creek District, Graham (1994) found evidence of pre-AD 600 settlements
along the coastal plain, though much of that region’s settlement history is Late and
Terminal Classic. Sites such as Pomona, Mayflower, and Kendal are located along rivers
seasonally navigable by canoe, and have been suggested to be interconnected nodes along
river systems (Graham 1994: 320). Coastal sites may have been organized as subsistence
bases that engaged in procurement of marine and estuarine resources (Graham 1994: 316)
or, in some cases, also mediated maritime trade networks (McKillop 2005a, 2005b).
Among the earliest sites in southern Belize is the coastal shell midden of Butterfly Wing
at the mouth of Deep River, which is thought to date to the Late Preclassic based on
sherds of mammiform tetrapod vessel supports and outflaring wall dishes (McKillop
1996:57, 2010:96). The presence of obsidian and other exotic goods identifies it as a
trading port, and links it with other Late Preclassic sites at Cancun, Cerros and Moho
Cay. Radiocarbon dates from Early Classic settlements on Wild Cane Cay indicate
maritime communities established by AD 300, though mercantile seafaring was largely a
post-AD 500 phenomenon that persisted into the Postclassic (McKillop 2005a, 2005b,
2006).
The early communities closest to Uxbenká were in the southeastern Petén
(Guatemala), positioned along the western foothills of the Maya Mountains. Most of
21
these settlements postdate AD 600, though there were Preclassic occupations at Sacul,
Ixkun, Xutilha, and Ixtonton in the Dolores area (Laporte 1994, 2001; Laporte and
Ramos 1998). Throughout the watersheds that drain the western Maya Mountains of
Guatemala, including the Rios Machaquila, San Luis, and Pusilhá, there is evidence of
continuity between the Preclassic and Early Classic in what Laporte (2001:17) called the
“Peripheral Chicanel” sphere, defined by the continuation of Preclassic ceramic types
well into the Early Classic period. Laporte suggested a geopolitical landscape of
competing rural elites autonomous from the larger central Petén polities from AD 100 to
AD 600 (Laporte 1996a, 1996b; Laporte and Ramos 1998). The southeastern Petén, like
southern Belize, witnessed greater population centralization during the Late and Terminal
Classic periods, and evidence for Early Classic occupations is spotty (Brady 1989: 207;
Laporte 2001).
The only other Preclassic or Early Classic complex polity known in the region is
Ek Xux, located in the interior of the eastern Maya Mountains along the Bladen Branch
of the Monkey River (Dunham and Prufer 1998). Nine sites with public architecture are
known from survey in the eastern flank of the Maya Mountains, but excavation data only
exist for Ek Xux and Muklebal Tzul, both located in adjacent valleys near the headwaters
of the Bladen Branch. Ceramic evidence suggests Ek Xux was settled during the Late
Preclassic and persisted as a relatively small community through the sixth century AD.
Muklebal Tzul, located on a series of high ridges 3 km to the west of Ek Xux, appeared
rather suddenly on the landscape after AD 600 and quickly eclipsed its small neighbor
(Prufer 2005a).
22
With the exception of Uxbenká and Ek Xux, southern Belize apparently hosted
few population centers through most of the Early Classic, until the region rapidly grew to
include at least 10 monument bearing polities and over 100 smaller communities after ca.
AD 550. The best known of these are Lubaantun, Pusilhá, and Nim Li Punit. Hammond
(1975:52) conducted excavations at Lubaantun and, based primarily on ceramics,
suggested that the site was founded between AD 679 and AD 783 (i.e., Maya calendar
date 9.15.0.0.0 +/- 1 katun). He also noted that the ceramic assemblage was dominated by
Tepeu 2/3 Petén styles of the Late Classic (maximally AD 700-890). Hammond argued
for links between southern Belize and sites in the Pasión River area of the western Petén
(1975: 295), which are supported by more recent studies at other Late Classic centers
(Braswell et al. 2005; Prufer 2005a; McKillop 2006). Lubaantun lacks epigraphic history
from monuments, though three carved ballcourt markers have been stylistically dated to
the Late Classic (Wanyerka 2004). Pusilhá was excavated by a British Museum
expedition (Joyce 1929; Joyce et al. 1927), Hammond (1975:274), Leventhal (1990,
1992) and Braswell (Bill and Braswell 2005; Braswell et al. 2004). Hieroglyphic texts
suggest that the polity may have formed as late as AD 570 and persisted at least through
AD 790. Excavations in core and domestic contexts support this chronology (Braswell
and Prufer 2009: 48), though small amounts of Early Classic materials have been
recovered from cave sites in the vicinity. Ceramic data suggest a Late Classic affiliation
closely aligned with Tepeu sphere polities in the Petén, particularly in the Pasión and
Petexbatun areas (Bill and Braswell 2005). Nim Li Punit is the least studied polity in the
region. It is located on a 100 m high ridge overlooking the coastal plain (Hammond et al.
1999). Most of the published chronological material on Nim Li Punit comes from 25
23
carved monuments found in the elite plazas of this highly consolidated center. These have
been interpreted to suggest the site was occupied only during the Late Classic, with stelae
erected between AD 711 and AD 830 bracketing a short dynastic history for the polity,
but the possibility of earlier and later non-dynastic site use must be kept open. The Nim
Li Punit inscriptions are described as both “unique and idiosyncratic” (Grube et al. 1999:
36) with examples of reverse order readings, inverted calendar signs, and evidence that
the placement and carving of the monuments may be temporally separated events.
Epigraphers have also suggested that the people of Nim Li Punit regularly interacted with
occupants of sites to the southeast, based largely on the presence of a possible toponym
glyph for Copán (Wanyerka 2009: 465).
Artifacts and monuments indicate ties between southern Belize and the central
Petén from AD 370-500, probably via trade routes through the southeastern Petén (Prufer
2005a). Epigraphic accounts of ties developing after AD 500 between southern Belize
and sites located in the southeast periphery have been proposed, e.g., with Copán and
Quirigua (Braswell et al. 2005; Grube et al. 1999; Marcus 1993; Wanyerka 2009: 440-
477) or Altun Ha (Wanyerka 2009: 473). Archaeological evidence to corroborate these
relationships remains to be found. By the 9th century AD there is little archaeological
evidence of any substantial inland Postclassic occupation, though the difficult work of
identifying and recovering these contexts in southern Belize has barely begun. The
persistence of maritime trade into the Postclassic Period at coastal sites suggests the
potential for a continued, if politically diminished, presence in the inland areas of
southern Belize.
24
Methods
Radiocarbon Sampling and Measurement
In a region with few absolute dates from archaeological contexts, the AMS
radiocarbon dating program allows the UAP to develop an independent chronology of the
growth and contraction of Uxbenká as a political center. Charcoal and other organic
samples from well-documented stratigraphic contexts (see below) were prepared along
with standards and backgrounds at the University of Oregon Archaeometry Facility and
the University of California Irvine Keck Carbon Cycle AMS Facility (UCI KCCAMS)
following standard practices as previously described by Prufer et al. (2011:Note 1).
Samples for dating were collected during excavations directly from discrete features (e.g.,
hearths, burn features), plaster floors, or from within construction fill. These were taken
“at the trowel’s edge”, not recovered from screened sediments. Where possible a single
piece of wood or charcoal was selected to avoid the averaging inherent in bulk samples,
and pieces likely to be shorter-lived (e.g., twigs) were chosen to reduce any old wood
effect (Schiffer 1986; Kennett et al. 2002). All dates are reported in Table 2.2 as
conventional radiocarbon ages corrected for fractionation with measured δ13C according
to Stuiver and Polach (1977). Calendar ages discussed in the text are 2-sigma calibrated
ranges (95.4% probability; for clarity, discontinuous ranges are simplified in the text).
Calibrations were produced using OxCal 3.01 (Bronk Ramsey 1995, 2001, 2009),
employing the IntCal09 atmospheric curve (Reimer et al. 2009). Calibrated dates are
discussed in terms of ‘cal AD’ or ‘cal BC’ as distinct from dates derived from epigraphic
and seriational methods.
25
Table 2.2. AMS 14C dates from Uxbenká used in Bayesian modeling Sequence/ Phase
UCIAMS-#
Provenience Conventional 14C age (BP)
2-σ cal range (prior)
Group A West A1 Sub Op 08-4 56360 Structure A1. Buried Structure Fill, 198cmbd. 1840±15 AD 120-230 56359 Structure A1. Level 5, 169cmbd. 1780±15 AD 140-200 (3.8%)
AD 210-330 (91.6%) 56367 Structure A1. Level 4, 108cmbd Fea. 1. 1635±15 AD 350-370 (1.2%)
AD 380-440 (88.3%) AD 480 530 (5.9%)
56368 Structure A1. Level 4, 120cmbd Fea. 2. 1585±15 AD 420-540 Group A A6 SubOp 07-3 & Plaza Plaster SubOp 07-5 46297 Structure A6. Level 5, 367cmbd. First fill. 1755±25 AD 220-390 42807 Structure A6. Level 5, 292 cmbd. Second fill. 1720±15 AD 250-390 42805 Structure A6. Level 5, 224 cmbd. Second fill. 1700±15 AD 250-300 (18.8%)
AD 320-410 (76.6%) 42809 Structure A1. Level 5, in plaza plaster floor. 1490±15 AD 540-610 46298 Structure A1. Level 5, in plaza plaster floor. 1585±25 AD 410-540 Group B SubOp 08-7 Unit 2 56361 Unit 2. Level 6 Construction Fill, 204 cmbd 1755±15 AD 235-340 56371 Unit 2. Level 6 Construction Fill, 143 cmbd 1735±15 AD 240-380 56370 Unit 2. Level 5 Construction Fill, 139 cmbd 1730±15 AD 250-390 56369 Unit 2. Level 5 Construction Fill, 121 cmbd 1760±15 AD 230-340 57044 Unit 2. Level 3. On Level 4 Floor, 95 cmbd 1745±15 AD 230-350 Group B Other 56362 Structure B2 SubOp 08-9. Base of wall. 1770±15 AD 210-340 56365 Structure B14 SubOp 08-10. Level 5A. 191
cmbd. 1725±15 AD 250-390
56364 Structure B1 SubOp 08-8. Base of staircase. 1315±15 AD 650-710 (78.3%) AD 740-770 (17.1%)
Group D Late Preclassic/Early Classic Phase 67955 SubOp 9-15 Unit 2. Level 3 Box Lu’um blw
plaster. 136cmbd 1830±15 AD 130-240
67238 SubOp 9-14 Unit 1. Level 7. 4th Floor Fill. 192cmbd
1775±20 AD 140-200 (4.6%) AD 210-340 (90.8%)
67961 SubOp 9-14 Unit 1. Level 7. 3rd Floor Fill. 169cmbd
1750±20 AD 230-350 (94.3%) AD 360-380 (1.1%)
67960 SubOp 9-14 Unit 1. Level 6. 2nd Floor Fill. 153cmbd
1800±20 AD 130-260 (90.8%) AD 300-320 (4.6%)
67959 SubOp 9-14 Unit 1. Buried Structure Fill. 158 cmbd
1710±15 AD 250- 300 (30.7%) AD 310- 400 (64.7%)
67239 SubOp 9-13 Structure 5. Level 4. 95 cmbd 1695±20 AD 250-300 (17.3%) AD 320-410 (78.1%)
Group D Late Classic Phase 67957 SubOp 9-14 Level 3 Box Lu’um. 105cmbd 1345±15 AD 650-685 67958 SubOp 9-14 Level 3 Box Lu’um. 80cmbd 1465±15 AD 565-640 67965 SubOp 9-13 Structure 5. Level 3 63 cmbd 1225±15 AD 710-750 (16.8%)
AD 760-880 (78.6%)
The architectural stratigraphy at Uxbenká is complex because most structures
have several construction phases and remodeling episodes, and a range of natural and
cultural site formation processes (see Schiffer 1987) have and continue to affect the
deposits. In some cases older materials may have been reused for the construction of later
structures. Interpretation is further complicated by post-depositional alterations at
Uxbenká, and most Lowland Maya sites, due to erosion, bioturbation by burrowing
26
animals and tree-throws, modern landuse and looting. Rosenswig (2009) provides a
cogent treatment of the often under-appreciated complexities involved in structural
stratigraphy in Mesoamerica, particularly at Classic Period sites where (arguably) more
focus is placed on the “glamour” of elaborate architecture than on the quotidian aspects
of formation processes (Rosenswig 2009:2, amplifying Shott 2006:4). Despite devoting
effort to careful excavation and stratigraphic correlation between observed architectural
elements, cross-referencing multiple individual radiocarbon sequences through common
features such as plaster floors is often difficult. Stratigraphic information recorded during
excavations in the 2006 to 2009 field seasons were used to select the sample of
radiocarbon dates that are incorporated into the Bayesian analysis of Groups A (the Stela
Plaza), B and D (Figure 2.1A). Emphasis was placed on excavation units exhibiting clear
natural and architectural stratigraphy, including plaster floors, masonry construction, and
multiple construction episodes.
The Bayesian Framework
Classical statistical analysis has dominated archaeological inquiry and is well
suited to a wide range of observations made by archaeologists (Drennan 2010; Shennan
1997; Thomas 1986). In contrast to classical statistics, Bayesian statistical analysis
derives posterior information (a posteriori) by combining prior information (a priori), a
likelihood function (a particular probability function) and the available data (Buck and
Millard 2004: p. VII). The best examples in archaeology come from chronology building
where a variety of non-quantitative contextual information (e.g., stratigraphic position,
27
Figure 2.1. Detail maps showing excavations at: A) the Uxbenká site core; B) Group A (Stela Plaza); C) Group B; and D) Group D (original figures by C. Ebert).
diagnostic artifact assemblages) can be integrated with probability distributions from
radiocarbon dates (Bayliss and Bronk Ramsey 2004; see below).
The major benefits of a Bayesian approach are that a statistical environment is
created that incorporates a wider range of information about stratigraphy and
archaeological materials, and that the results of these models can be used to direct
28
research and make sampling decisions. Using a priori information can make some
researchers uneasy (see Steier and Rom 2000), but by forcing the assumptions of the
priors to be made explicit it provides a framework to formalize assumptions and to build
and test multiple models with new data. Agreement indices (A) provide a way of
determining how each alternative model fits with the available data, and are generated for
the posterior distributions of each radiocarbon date in a model, as well as the overall
model itself (Bronk Ramsey 2000: 201). Agreement indices falling below a critical value
(A′c = 60%) indicate a poor fit of data with the model, and can be used to identify
potential outlier dates or problematic stratigraphic assumptions in the model. It should be
noted that, strictly speaking, when A>A’c (i.e., there is agreement between the model
structure and the dates) it does not mean that the model assumptions and structure are
correct: it simply tells us that we have no reason based on the data at hand to reject the
model as it stands.
A list of OxCal commands and the relevant archaeological phenomena that are
commonly encountered during excavation are presented in Table 2.3. The reader is
referred to the OxCal’s supporting documentation for detailed considerations of analysis
and command structures, as well as other published archaeological case studies in Britain
(Bayliss et al. 1999, 2007 [and articles therein]), the Mediterranean (Bronk Ramsey et al.
2010; Manning et al. 2006), and Mexico (Kennett et al. 2011).
29
Table 2.3. Examples of OxCal commands and relevant stratigraphic situations OxCal Command Stratigraphic Situation Phase (Unordered Group)
Multiple dates within a fill.
Multiple features on a living surface occupied for some duration Groups of dates separated by a common stratigraphic marker e.g., a floor, a sterile sediment layer, tephra, or distinctive ceramic assemblages.
Sequence (Ordered Group)
Dates separated by a series of plaster floors. Dates on materials in well stratified middens.
Series of phases Boundaries Events that bracket the beginning and end of a phase but are not directly dated, e.g.,
excavation of a burial or storage pit; clearing or leveling a site before construction; cessation of construction; partial demolition of a structure.
Event An undated event not necessarily related to a phase, thus differing from boundary in that it
could be within a sequence. Cross-reference When a common stratigraphic marker can be correlated between two or more sequences,
e.g., a layer of pavers, a floor, a tephra, or a burning event, can be traced between sequence with otherwise unconnected profiles.
Span Calculates the span of time represented by the elements of a phase, e.g., how long a living
surface was used before being covered over or replastered.
A sample of 28 AMS radiocarbon dates from the 2006 through 2009 field seasons
was included in this analysis. Radiocarbon data for samples from Groups A and B have
been reported and discussed in Prufer et al. (2011); dates and stratigraphic information
for Group D are drawn from the Uxbenká Archaeological Project technical report on the
2009 excavation season (Ebert et al. 2010).
Results
Group A (Stela Plaza)
Group A, also known as the Stela Plaza, is a plaza group set on a hilltop in the
eastern part of the Uxbenká site core containing six known structures and 23 recorded
stelae (Figure 2.1B). Leventhal worked at Group A in the late 1980s, recording and
describing the stelae and conducting excavations in the plaza itself. Dates preserved on
six of the monuments indicate monument production and dedication occurred during the
Early and Late Classic periods (AD 378 to AD 781; Table 2.4), with the dates of AD 378
30
stylistically attributed to Stela 11 and a calendar round date in AD 455 on Stela 23
making these the earliest datable monuments in southern Belize.
Table 2.4. Stela dates from Group A, Uxbenká Monument Long Count Date Gregorian Date Comments
Stela 11 - AD 378 After the reign of Chak Tok Ich’aak I; Schele and Looper (1996) suggest AD 437 for this stela
Stela 23 09.01.00.00.00 AD 455 Period ending date derived from a calendar round date (Prufer and Wanyerka 2005)
Stela 14 09.12.00.00.00 AD 672-692 Partial inscription, inferred 12th katun Stela 19 09.12.11.13.11 AD 684 Stela 22 09.16.00.00.00 AD 751 Period ending date Stela 15 09.17.10.00.00 AD 781 Period ending date
The Uxbenká Archaeological Project team excavated several structures in the
Stela Plaza, including A1 (the largest construction in the group), A4, A5 and A6.
Multiple test trenches were also excavated across the plaza floor between 2006 and 2010.
Results of these investigations suggest that the hilltop was leveled in the latest part of the
Late Preclassic, with some of the earliest construction fills below structure A1 dating to
cal AD 120-230 (UCIAMS-56360). Evidence of walls and other structural features in
direct contact with the mudstone bedrock (known as nib in the local Mopan Maya) under
A1 and in front of A6 indicates that sections of the plaza must have been completely
excavated to bedrock before major construction of Group A took place (Prufer et al.
2011). A date on charcoal below the A6 wall is also consistent with a Late Preclassic
clearing event (cal AD 130-330; UCIAMS-33400).
Excavations along the margin of A1 (specifically SubOps 07-5 and 08-4) reveal
multiple phases of construction and remodeling related to periodic reorganization of the
plaza for ceremonial or political purposes (Prufer et al. 2011). After the initial Late
Preclassic clearing event, it appears that a much smaller structure was put in place under
what is now the west flank of Structure A1. A portion of one of the walls of this structure
31
was uncovered at ~180 cm below the surface of A1 in SubOp 08-4 Unit 1, measuring
roughly 1 m high and made of 10-12 courses of the local sandstone slabs typically used to
build these structures (Figure 2.2). The Late Preclassic 14C date noted above (UCIAMS-
56360) was recovered from construction fill within this buried structure. The early
structure was built over a layer of crushed nib fill directly above bedrock. In contrast to
the rest of the known architecture in the Stela Plaza, which is oriented roughly along (or
just east of) the cardinal axes, the wall exposed in Unit 1 is oriented at 53°/233°mN. A
fill deposit consisting of sediment and loose sandstone slabs covers the buried structure
and contains one Late Preclassic/Early Classic charcoal date of cal AD 140-330
(UCIAMS-56359). Two burn events occur on top of this fill deposit and suggest a
persistent surface dating later in the Early Classic (Feature 1: UCIAMS-56367, cal AD
350-530; Feature 2: UCIAMS-56368, cal AD 420-540). Roughly 1 m of subsequent
construction fill overlies these features and presumably represents renewed building on
Structure A1 at or after the end of the Early Classic.
A 6m-long profile exposed from the Stela Plaza floor into the eastern side of
Structure A1 in SubOp 07-5 shows the stratigraphic relationship of the plaza construction
to the later additions to the building (Figure 2.3). Excavation into the eastern flank of
Structure A1 (on a flat platform similar to the one where SubOp 08-4 was placed) cut
through a mixed layer of overburden and sandstone blocks and two layers of crushed
bedrock fill before revealing a burn feature dating to the Late Preclassic Period
(UCIAMS-42825; cal AD 70-220) and a charcoal sample from a deeper deposit of dark
soil and burned ceramics dating to the Early Classic (UCIAMS-42808; cal AD 250-390)
32
Figure 2.2. Profile of Unit 1, SubOp 08-4 on Str. A1 showing location of AMS 14C samples and modeled calibrations.
overlying the nib bedrock. The inconsistency between the dates could be due to
disturbance related to later construction, or an old wood effect in the charcoal from
Feature 1. Assuming that the lower deposit is accurately dated, this indicates a surface
that had originally been exposed in the beginning of the Early Classic Period and was
subsequently buried by construction of the later facade of Str. A1. The stepped facade of
Str. A1 is exposed in the same profile, where collapse debris was removed to reveal the
remaining intact south face of the building. This wall was built directly on the nib
bedrock, indicating that any overlying soil in what would become the plaza floor was
removed before this time. As described by Prufer et al. (2011) the depths where bedrock
is encountered differ by ~1.5 m on Structure A1 and the plaza floor suggesting a sharp
discontinuity behind the facade. This could be due to a natural joint in the bedrock, as has
33
Figure 2.3. Profile of SubOp 07-5 on Str. A1 showing location of AMS 14C samples and modeled calibrations.
been observed in geomorphic excavations and stream channels in the Uxbenká site core
and elsewhere, or a purposeful modification of the bedrock by the ancient occupants to
take advantage of an existing hilltop feature to create a more imposing ceremonial
structure. Regardless, clearing down to bedrock, erecting the facade, and burying the
earlier architecture was a key event in the development of the Stela Plaza, whose date can
be constrained by a thick plaster floor in the plaza that abuts and therefore post-dates the
facade. Two dates were obtained from charcoal recovered from within the plaster floor,
which likely represent material incorporated during the plaza’s construction and use
during the latest part of the Early Classic (UCIAMS-46298: cal AD 410-540; UCIAMS-
42809: cal AD 540-610).
34
A third series of dates relates to the smaller structure A6 on the east side of the
Stela Plaza excavated in 2006 (SubOp 06-7) and 2007 (SubOp 07-3). SubOp 06-7 was
conducted off the structure and revealed a stone wall in front of Str. A6 that is inferred,
based on its location and alignment, to be part of an earlier construction that may have
been leveled or simply buried during the construction of A6. The wall sits directly on
bedrock, and a date on charcoal from beneath the wall straddles the Late Preclassic/Early
Classic transition (UCIAMS-33400; cal AD 130-330). Str. A6 itself appears to have been
constructed in at least three phases as indicated by a series of fill layers capped by plaster
floors (Figure 2.4). Three AMS dates on charcoal put these construction events in the
Early Classic, with the earliest layer, which sits upon bedrock, dating to cal AD 220-390
(UCIAMS-46297). Two dates from the second fill layer fall into a similar timeframe
(UCIAMS-42807: cal AD 250-390; UCIAMS-42805: cal AD 250-410).
Each of the profiles in these excavations can be modeled as a separate sequence.
This allows for timing of events that are not directly dated to be estimated, such as
clearing the plaza to bedrock or constructing a facade. For SubOp 08-4, the sequence
begins with a boundary, the earliest use of the hilltop, followed by the construction of the
early structure under Structure A1, the burial of that structure (an event), followed by the
creation of the two burned features, and ending with the final boundary, the end of
construction of Str. A1. In SubOp 07-5, the sequence begins with a terminus post quem,
which is a cross-reference to the date in SubOp 08-4 on the fill above the buried
structure, UCIAMS-56395, since it is assumed that all of the construction exposed on the
east side of Str. A1 post-dates the earlier construction. This is followed by the use of the
burned ceramic layer within Str. A1. The Late Preclassic date on Feature 1 (UCIAMS-
35
Figure 2.4. Profile of SubOp 07-3 on Str. A6 showing location of AMS 14C samples and modeled calibrations
42825), as mentioned above, is problematic when included even as part of an unordered
group (a phase) with the later date on the deeper ceramic layer. Models that include this
date produce very low agreement indices and so it is excluded from this sequence. A
boundary representing the placement of nib fill and the construction of the Str. A1 facade
follows, and the two dates from within the thick plaster floor in front of A1 are modeled
as a phase followed by a final boundary that represents subsequent deposition above that
floor. The sequence for Str. A6 cannot be stratigraphically linked to those in A1 based on
current knowledge. It begins with a boundary, the clearing of that section of the plaza
down to bedrock, then the construction of the wall in front of Str. A6, the placement of
the first fill in A6, followed by a phase comprising the two dates in the second fill event,
and ending with a boundary representing the placement of the third fill layer. Modeled
results for these sequences are presented in Table 2.5.
36
Table 2.5. Modeled results for three Group A stratigraphic sequences Sequence UCIAMS-
# Provenience Conventional
14C age (BP) Modeled 2-σ cal range
West A1 08-4 Boundary Earliest Group A 50 BC - AD
220 56360 Str. A1. SubOp 08-4. Buried Structure Fill, 198cmbd. 1840±15 AD 120-230 Event Burial of Early Structure AD 150-310 56359 Str. A1. SubOp 08-4. Level 5, 169cmbd. 1780±15 AD 220-330 56367 Str. A1. SubOp 08-4. Level 4, 108cmbd Fea. 1. 1635±15 AD 355-440 56368 Str. A1. SubOp 08-4. Level 4, 120cmbd Fea. 2. 1585±15 AD 420-540 Boundary End of Early Classic Construction AD 430-700 East A1 SubOp 07-5 TPQ UCIAMS-56359 (cross-referenced) 42808 Str. A1. SubOp 07-5. 238N/-20E. L.7, burned layer. 1725±15 AD 250-390 Boundary Placement of Nib Fill/Construction of Facade AD 310-540 46298 Str. A1 .SubOp 07-5. 236N/-20E. L.5, in plaster floor. 1585±25 AD 420-550 42809 Str. A1. SubOp 07-5. 236N/-20E. L.5, in plaster floor. 1490±15 AD 540-610 Boundary Deposition Above Plaza Plaster Floor AD 550-770 Str. A6 SubOps 06-7 & 07-3 Boundary Clearing to Bedrock AD 20-320 33400 West of Str. A6. SubOp 06-7. Level 4, beneath wall. 1790±25 AD 160-330 46297 Str. A6. SubOp 07-3. Level 5, 367cmbd. First fill. 1755±25 AD 230-340 42807 Str. A6. SubOp 07-3. Level 5, 292 cmbd. Second fill. 1720±15 AD 250-390 42805 Str. A6. SubOp 07-3. Level 5, 224 cmbd. Second fill. 1700±15 AD 250-400 Boundary Placement of Third Fill AD 260-520
Group B
Group B was first identified by Hammond (1975:289-290) and later excavated by
Leventhal (1992:145) who designated it as the North Group. It consists of an enclosed
plaza on a hilltop at the northern end of a 400m-long modified ridge to the west of Group
A. The main structures include a temple (Str. B1), a ballcourt (Str. B6 and B7) and three
patio structures (Str. B3, B5, and B11; Figure 2.1C). The UAP excavations that provide
the data for this analysis were conducted in 2008. Excavations of the front stairway of
Str. B1 (Op 08-8) produced a Late Classic ceramic assemblage consistent with elite ritual
use, including numerous unslipped modeled effigy censer fragments, Petén Gloss Wares
and other polychrome ceramics. A single AMS date from under a slumped step produced
an age range of cal AD 650-710 (78.3%) and AD 740-770 (17.1%) (UCIAMS-56364).
An Early Classic component at Group B is evident in the excavations in the main plaza
37
and a smaller bench to the west side of Str. B1, all buried by later construction. Units
placed between Strs. B2 and B3 (Subop 08-9) uncovered a section of a 1.6 m high
masonry wall buried below the visible structures, and not showing any clear connection
to the later architecture in terms of layout or organization. A single radiocarbon sample
from the base of the wall dated to cal AD 210-340 (UCIAMS-56362), which is consistent
with Early Classic construction in the Stela Plaza. A charcoal date from a buried midden-
like fill stratum in Str. B14 also falls into the Early Classic at cal AD 250-390 (UCIAMS-
56365).
The episodic nature of Early Classic construction in Group B is revealed in
excavations in front of Str. B9, a low platform on the southwestern edge of the plaza
(SubOp 08-7). Three construction episodes are marked by plaster floors and structural
elements exposed in Unit 2 (Figure 2.5). A series of large cut limestone and sandstone
blocks were found lying on the nib bedrock in the basal deposits of this unit (2 mbs).
These blocks were probably put into place to level and extend the southwest edge of the
plaza after clearing the space down to bedrock. A fill layer containing ceramic sherds and
river snail (jute; Pachychilus sp.) shells overlies the bedrock and abuts the block
construction, and is capped by a thin plaster floor. Two charcoal samples from this
stratum date to cal AD 235-340 (UCIAMS-56361) and cal AD 240-380 (UCIAMS-
56371). What appears to be collapsed rubble from a constructed wall overlies this floor
and is covered by another layer of fill and a second plaster floor. Two charcoal samples
from this fill date to cal AD 230-340 (UCIAMS-56369) and cal AD 250-390 (UCIAMS-
56370). Finally, a third fill and plaster floor is exposed immediately below the modern
38
Figure 2.5. Profile of Unit 2, SubOp 08-7 in Group B showing location of AMS 14C samples and modeled calibrations
surface (A horizon) of the plaza floor. Charcoal recovered from directly on top of the
plaster floor also dates to the Early Classic at cal AD 230-350 (UCIAMS-57044).
With the exception of the five dates within Unit 2, SubOp 08-7, the stratigraphic
relationships of the Group B AMS 14C dates are difficult to establish with certainty. This
is primarily because the individual suboperations are widely separated across (and off of)
the plaza. However, to make use of all the existing data I incorporate them all into a
broad sequence, with the Late Preclassic/Early Classic dates organized as a phase,
followed by the single Late Classic date from the slumped step in front of Str. B1. Three
boundaries are established within this sequence: the earliest construction and clearing
activities on Group B; the transition between Early and Late Classic construction
activities (e.g., the Str. B1 staircase and presumably the ballcourt); and the latest
construction activities in the Late Classic. Within the Late Preclassic/Early Classic phase,
39
the SubOp 08-7 dates are placed in a sequence beginning with a boundary representing
the construction of the wall, followed by two phases containing the pairs of dates from
the L6 and L5 fills, and the final date on charcoal above the upper most floor. These are
separated by boundaries representing the construction and use of the three floors in the
sequence. As noted the two dates from between Strs. B2 and B3 and from Str. B14 are
included with this sequence in an unordered phase. Modeled results for the sequence are
presented in Table 2.6.
Table 2.6. Modeled results for the Group B stratigraphic sequence Sequen ce/ Phase
UCIAMS-#
Provenience Conventional 14C age (BP)
Modeled 2-σ cal range
Boundary Earliest Group B AD 60-310 Grp B 08-7 Unit 2 Boundary First Wall Constructed AD 210-320 56361 Level 6 Construction Fill, 204 cmbd 1755±15 AD 240-320 56371 Level 6 Construction Fill, 143 cmbd 1735±15 AD 245-320 Boundary Plaster Floor between L5/L6 AD 250-325 Difference First Wall constructed - Floor between L5/L6 -5-75 cal yr 56370 Level 5 Construction Fill, 139 cmbd 1730±15 AD 255-335 56369 Level 5 Construction Fill, 121 cmbd 1760±15 AD 250-335 Boundary Level 4 Plaster Floor AD 260-345 Difference Floor between L5/L6 – L4 Floor -5-50 cal yr 57044 Level 3. On Level 4 Floor, 95 cmbd 1745±15 AD 270-335 Boundary Level 2 Plaster Floor AD 270-420 Difference L4 Floor – L2 Floor -5-95 cal yr Grp B Early Classic 56362 Between Str. B2-B3 SubOp 08-9. Base of wall. 1770±15 AD 230-340 56365 Str. B14 SubOp 08-10. Level 5A. 191 cmbd. 1725±15 AD 250-380 Boundary Transition between Early and Late Classic Construction AD 290-670 Grp B Late Classic 56364 Str. B1 SubOp 08-8. Base of staircase. 1315±15 AD 650-770 Boundary Latest Group B AD 650-930
Three instances of the difference command are also included in the model to
estimate the duration between construction events in the Unit 2 sequence (the first
construction and the subsequent placement of plaster floors). The maximum 2σ ranges
for these estimates vary from 50 to 95 cal years, but the distributions are skewed towards
larger values, so the intervals between construction events may be much shorter, perhaps
every 15-25 years. Weighted means for these probability distributions suggest: ~23 years
40
passed between the construction of the wall and the placement of the plaster floor
between L5 and L6; ~16 years elapsed before the placement of the L4 floor; and the L2
floor was laid down ~27 years after that.
The wide estimated range for the boundary between the Early and Late Classic
construction phases (cal AD 290-670) is due to the lack of dates falling in the later part of
the Early Classic. It seems unlikely that there was no construction or modification of
Group B architecture during this period. However, it may simply reflect the areas
excavated and sampled during the 2008 excavations. The result indicates a chronological
issue to be addressed by ongoing strategic excavations at Group B.
Group D
Group D is located on the same long ridge as Group B and is immediately south
of Group C, which is contiguous with both (Figure 2.1D). Group D is conspicuously flat
as the result of leveling during Uxbenká’s construction. Primary structures include a ball
court that was subject to limited investigations in 2006 (Prufer et al. 2007) and a raised
open plaza surrounded by low (30-40 cm tall) walls and a few small platforms that were
excavated in 2009 (Ebert et al. 2010). This open plaza occupies roughly two-thirds of a
finger ridge that extends off the main landform to the east. Excavations revealed a series
of construction episodes and provided the sample of radiocarbon dates analyzed here.
Two 6x1 m stratigraphic trenches (SubOps 09-12 and 09-15) in Group D cut
through multiple fill and plaster layers within the plaza and indicated the broad outline of
construction events, while other excavations focused on Structure 5 (SubOp 09-13) and
the area immediately in front of the structure (SubOp 09-14). The generalized
41
stratigraphic sequence for the Group D plaza suggests that after initial clearing of the
ridgeline a fill of crushed nib bedrock was laid down to level the surface and then
plastered. A charcoal date recovered from within this nib fill in SubOp 09-15 Unit 2
(UCIAMS-67955) dates this event at the end of the Late Preclassic at cal AD 130-240.
Multiple fill and plastering episodes covered this initial building phase. Three plaster
floors were identified in Unit 2 and four were identified in SubOp 09-14 Unit 1 across the
plaza (Figure 2.6). Considering the differing number of floors in each unit and the
distance between them it is not possible to directly correlate these plastering events.
Multiple charcoal samples recovered from the Unit 1 floor fills promised to generate a
very detailed construction chronology for the plaza, but despite the apparently well-
stratified exposure, several reversals occur. Working from the stratigraphy, it appears that
at some point after the plaza was established, a now-buried structure was constructed in
the Early Classic. A single AMS 14C date from fill within this structure dates to cal AD
250-400 (UCIAMS-67959). Two plaster floors abutting this were constructed
subsequently and finally the entire structure was buried and plastered over completely.
Dates within these floor fills are problematic, though they all fall in the Early Classic. In
stratigraphic order the three fills date to cal AD 140-340 (UCIAMS-67238), cal AD 230-
380 (UCIAMS-67961), cal AD 130-260 (90.8%) and cal AD 300-320 (4.6%; UCIAMS-
67960). The three floors make a reasonable sequence on their own, but including the
structure fill date before them results in a very low agreement index (A=55.4%). In the
present case there is no clear justification for rejecting any one of these dates, though
bioturbation, old charcoal incorporated in the fill, and other processes are likely at work.
For the purposes of this analysis the dates were grouped as an unordered phase
42
Figure 2.6. Profile of Unit 1, SubOp 09-14 in the Group D plaza showing locations of AMS 14C samples and modeled calibrations
representing Early Classic construction. Above the highest plaster floor in the unit is a
distinctive stratum of dark midden-like soil containing ceramic sherds and capped by a
layer of sandstone slabs, presumably paving stones. Two charcoal dates from within this
fill fall securely in the Late Classic (UCIAMS-67958: cal AD 565-640; UCIAMS-67957:
cal AD 650-685), suggesting a renewal of construction on the plaza at this time.
The wall surrounding the plaza was exposed in SubOp 09-15 Unit 2 and runs into
Structure 5, which was excavated as part of SubOp 09-13. As it sits directly above the
plaster floors in Unit 2, it is inferred that after the series of Early Classic plastering
events, the low wall (60-80 cm high) was built around the plaza perimeter. Two dates
from 09-13 help to bracket the date of construction: cal AD 140-340 (UCIAMS-67239)
from Level 4 below the Structure 5 masonry; and cal AD 710-880 (UCIAMS-67965) in
the fill of the structure itself. The construction of the plaza wall can be used as a
boundary between two unordered phases. The Early Classic phase includes the AMS 14C
dates from the floor fills in SupOps 09-14 and 09-15, plus the Level 4 date from 09-
43
13.The Late Classic phase comprises the two dates on the paver fill from SubOp 09-14
and the structure fill date from 09-13. Two further events can be included as boundaries
in the overall Group D sequence: the placement of the pavers, and the subsequent
deposition of the surface scatters. Modeled results for Group D are presented in Table
2.7.
Table 2.7. Modeled results for the Group D stratigraphic sequence Sequence/ Phase
UCIAMS-#
Provenience Conventional 14C age (BP)
Modeled 2-σ cal range
Boundary Hilltop Cleared and Leveled AD 20-240 Grp D Early Classic Phase 67955 Grp. D. SubOp 9-15 Unit 2. Level 3 Box Lu’um blw
plaster. 136cmbd 1830±15 AD 130-240
67238 Grp. D. SubOp 9-14 Unit 1. Level 7. 4th Floor Fill. 192cmbd
1775±20 AD 170-340
67961 Grp. D. SubOp 9-14 Unit 1. Level 7. 3rd Floor Fill. 169cmbd
1750±20 AD 230-350
67960 Grp. D. SubOp 9-14 Unit 1. Level 6. 2nd Floor Fill. 153cmbd
1800±20 AD 130-320
67959 Grp. D. SubOp 9-14 Unit 1. Buried Structure Fill. 158 cmbd
1710±15 AD 250-390
67239 Grp. D. SubOp 9-13 Structure 5. Level 4. 95 cmbd 1695±20 AD 250-410 Boundary Plaza Wall Construction AD 270-580 Grp D Late Classic Phase 67957 Grp. D. SubOp 9-14 Level 3 Box Lu’um. 105cmbd 1345±15 AD 645-685 67958 Grp. D. SubOp 9-14 Level 3 Box Lu’um. 80cmbd 1465±15 AD 565-640 67965 Grp. D. SubOp 9-13 Structure 5. Level 3 63 cmbd 1225±15 AD 690-870 Boundary Pavers Placed AD 650-1300 Boundary Surface Scatters Deposited AD 700-
present
Though not well constrained, the model suggests that initial clearing and leveling
occurred at the end of the Late Preclassic at cal AD 20-240. A series of plastering
episodes and the construction of the buried structure in SubOp 09-14 followed, possibly
straddling the Preclassic/Classic transition and continuing into the Early Classic. The
construction of the plaza wall is unfortunately poorly constrained to cal AD 270-580, but
this range does place the event squarely in the Early Classic rather than the Late Classic.
The placement of the areally extensive paver layer is broadly estimated at cal AD 650-
1300 (with a 1σ range of cal AD 730-960), and the surface scatters must have been
44
deposited some time thereafter. The terminal ages of both the upper boundaries are
poorly constrained by this model, which would benefit from additional research. For
example, if diagnostic ceramics in the surface scatters indicated a distinctly Late Classic
component, a terminus ante quem could be added to the model at the assumed date of the
end of the Late Classic (i.e., at AD 800, following Demarest 2004). This points to a
direction for future chronological work at Uxbenká.
Discussion
Integration of the stratigraphic data with the existing high-resolution AMS 14C
dates from the urban core of Uxbenká provides strong evidence for its organization as a
sociopolitical entity during the Late Preclassic, with further bursts of architectural
modification at the beginning of the Early Classic and Late Classic periods respectively
(Figure 2.7). Initial clearing and leveling of the ridgeline hilltops that make up the civic-
ceremonial core began at Group A (the Stela Plaza) at cal 50 BC-AD 220, followed by
Group D at cal AD 20-240, and Group B only slightly later at cal AD 60-310. Accretion
of multiple plaster floors in each plaza group occurred across the transition from the Late
Preclassic to the Early Classic from ~ AD 200-400, a practice that appears to have ended
by cal AD 400 at Groups B and D. The only remodeling or construction evident in the
latter part of the Early Classic Period (between cal AD 400-550) appears to be the
addition of the facade construction on Structure A1 in the Stela Plaza that is estimated to
have been placed at cal AD 310-540. Estimates of the latest episode of construction at
each group are poorly constrained and provide little insight into the timing of the ultimate
demise of Uxbenká. Excavations targeting potential Late Classic and Terminal Classic
45
Figure 2.7. Summary of modeled calibrations for key construction episodes at Groups A, B and D.
contexts at Groups A and B are ongoing and may provide more concrete data to refine
these sequences. The clearest Late Classic event at Group A is the major plastering
episode of the plaza floor in front of Str.A1, estimated between cal AD 550-770.
Dedicatory dates on stela from Group A indicate monument carving had begun by the
Early Classic (St. 11, ca. AD 378; St. 23; AD 455) and continued into the Late Classic
after the last major plastering episode in the Stela Plaza (St. 22, AD 751; St.15, AD 781),
after which there is no secure radiocarbon evidence for use of the area.
46
The flurry of construction and replastering during the Late Preclassic and Early
Classic periods at Uxbenká is striking because it precedes the earliest dated monuments
in Group A by as much as 200-300 years. Similar bursts of remodeling and construction
activities are seen elsewhere in the Late Preclassic among the Lowland Maya. At San
Estevan in northern Belize, Rosenswig and Kennett (2008) describe a series of Late
Preclassic plastering episodes that cap Middle Preclassic midden layers and define what
would become the site center in the Late Preclassic. A direct AMS 14C date on charcoal
dates the later floor to cal 50 BC - AD 40 (UCIAMS-17903), which places the
construction of the first ballcourt at San Estevan in the Late Preclassic Chicanel ceramic
phase. Around the same time in the New River valley, multiple construction and
plastering episodes occurred through the Late Preclassic at Cuello (Hammond 1991;
Hammond and Gerhardt 1990), and monumental construction began at Lamanai
(Pendergast 1981) and Cerros (Scarborough 1983; Freidel 1986). The end of the Late
Preclassic (~AD 250-300) also witnessed the abandonment of El Mirador and Nakbe in
the northern Petén with the possibility of increased warfare and inter-polity conflict in
that region and highlights the localized factors affecting political development and
disintegration (Hansen 1998, 2006).
This analysis pushes the political integration of Uxbenká slightly earlier than
previously estimated by Prufer et al. (2011:218), and further removes the timing of the
initial large-scale landscape modifications from the potential Tikal connection inferred
from the mention of Chak Tok Ich’aak I on Stela 23 (AD 360-378). By that time, it
appears that the early major construction activities had ceased, though both Early Classic
stelae (11 and 23) at Group A do fall into the period when the outer plaza wall at Group
47
D was constructed, i.e., broadly estimated at cal AD 270-590. It is possible that the
remodeling at Group D had more to do with the site reorganization at the time of the
emergence of monument dedications at Uxbenká than do the earlier clearing and leveling
episodes. Since the perimeter wall along with the fills it contains obscure the earlier
features of that plaza, they may represent an effort to renew or rededicate that portion of
the site towards a new purpose, as Prufer et al. (2011) have argued for a buried Late
Preclassic settlement mound between Groups A and B.
I have been able to model construction episodes at Group A and B that fall into
the first part of the Late Classic. The extensive plastering episode at the Stela Plaza
occurred at cal AD 550-770, and the staircase construction and dedication on Structure
B1 is estimated at cal AD 650-770. The events at Groups A and B represent substantial
inputs of time and labor, and in the case of the staircase on Structure B1, great ceremonial
import. Not only do these estimated dates for these events coincide with the Late Classic
stela (ca. AD 672-781) at Uxbenká, they occur as the other major polities in southern
Belize appear (Lubaantun, and Nim Li Punit) or expand (Pusilhá). The presence of Tepeu
2/3 ceramics associated with this florescence also suggests more interaction with the
Petén during the Late Classic. Given that context, it is possible that the Late Classic
renovations at Groups A and B, presumably the respective ceremonial and civic centers
of the Uxbenká urban core, may have been an effort by local leaders to renew or
reinforce their position within a landscape of increasing sociopolitical complexity and
interaction regionally.
The stela dates suggest a point of caution in interpreting the results of this
analysis. Fewer construction episodes in the urban core during the later part of the Early
48
Classic and Late Classic Periods do not necessarily reflect a hiatus in the occupation and
use of Uxbenká during these periods. This analysis focuses primarily on architectural
events because of the stratigraphic constraints they provide for Bayesian modeling, so
more mundane or ritual activities that occurred between remodeling episodes are
underrepresented. Further, the latest occupations at any archaeological site are
stratigraphically shallowest and often the most disturbed deposits, which will preclude
them from this type of analysis (see Webster et al. 2004 for an in-depth treatment of this
problem at Copán). Most importantly in this regard, the present focus on the site core is at
the expense of the broader settlement history away from the urban core. In the case of
Copán, elite residences in the Copán pocket and more rural zones persisted for at least a
century after the Late Classic dynastic collapse (~AD 810) and some rural farming
populations persisted until sometime in the 11th century AD (Webster and Freter 1990;
Webster et al. 2004). So at Uxbenká the few construction events known during the later
part of the Early Classic (i.e., ca. AD 400-600) may not be representative of a “hiatus” at
the polity on the larger scale, but merely a shift in focus to other residential communities
outside of the site core. Work underway at Uxbenká in elite residential groups near the
urban core and others in more rural agricultural settlements should provide an interesting
test of these ideas.
The attempt to integrate a large number of high-resolution AMS 14C dates with
stratigraphic information within a Bayesian framework at Uxbenká provides a model for
applying this approach to other stratigraphically complex Mesoamerican sites. The
demands on the quality of archaeological information and the dated contexts are quite
high, and the proper interpretation of stratigraphic associations is crucial. Using a
49
Bayesian dating approach forces consideration of excavation strategy and sampling
techniques before excavations begin, and ideally to use insights gained in one season as a
priori data to guide excavations and 14C sample collection in subsequent seasons. In the
case of Uxbenká, there is now a better understanding of its early construction history. The
use of OxCal to estimate events that are not directly datable has pushed the establishment
of this polity back earlier than previously thought (Prufer et al. 2011). On the other hand,
the poorly constrained events within the Late Classic and Terminal Classic construction
sequences have crystallized numerous issues involved in dating those periods that are key
to understanding the processes of political disintegration in the tropical Maya lowlands.
Using this current knowledge, it is possible to take strategic aim at the parts of the site
most likely to contain the more elusive later construction phases in an efficient and
focused manner.
Conclusions
The Bayesian chronology developed here provides new insights into the
developmental history of Uxbenká’s urban core and provides a statistical framework for
future chronological refinement. The earliest leveling and clearing at Group A (the Stela
Plaza) began during the Late Preclassic at cal 50 BC – AD 220, roughly 100-200 years
earlier than previously thought (Prufer et al. 2011). This was followed by similar
landscape modifications at Group D (cal AD 20-240) and Group B (cal AD 60-310) and a
period of multiple plastering and remodeling episodes in both plazas. The leveling and
construction during the Late Preclassic and the Early Classic that established the nascent
urban core of Uxbenká preceded all evidence for dated stone monuments at the site, as
50
the earliest known stela was dedicated in AD 378. Based on the available evidence there
is relatively little construction in the site core that dates after the Early Classic Period
from ca. AD 400-600. However, the Group A plaza was substantially replastered in the
Late Classic at cal AD 550-770 along with the construction and dedication of a staircase
in Group B (Structure B1; cal AD 650-770). These events coincide with the dedication of
stela at Uxbenká and the appearance or expansion of other regional polities (e.g., Pusilhá,
Lubaantun, Nim Li Punit) that is possibly tied to increased interaction with the Petén
region. Secure Terminal Classic contexts have been difficult to identify, but remain a
focus of ongoing investigations at Uxbenká.
51
CHAPTER III
CHANGING AGRICULTURAL AND URBAN LANDSCAPES AT THE CLASSIC
MAYA CENTER OF UXBENKÁ, BELIZE
The work presented in this chapter was developed as an unpublished co-authored
manuscript with Dr. Keith M. Prufer and Dr. Douglas J. Kennett. I conducted the
geoarchaeological excavations at Uxbenká, recorded the stratigraphy, processed the
radiocarbon samples reported here, and analyzed the data. Fieldwork was conducted
under the supervision of Dr. Prufer. Dr. Kennett provided useful suggestions on the
integration of climate and geomorphic records, and valuable interpretations of the
possible land use strategies at Uxbenká.
Contemporary problems of deforestation and erosion have become synonymous
with the expansion of nation-states, global population increases, and intensified
agricultural production. This has stimulated archaeologists to consider landscape
transformation and the environmental impacts of agricultural systems (Barker 2008;
Bellwood 2005; Diamond and Bellwood 2003; Kennett and Winterhalder 2006; Smith
2007) and their expansion associated with the proliferation of state level societies during
the last 6000 years (Dunning et al. 2002; Kolata 1986; O’Hara et al. 1993; Redman 1992,
1999; Zeder 1991). Virtually all models of sociopolitical development and collapse
consider landscape transformation and associated decreases in yields, agricultural or
otherwise, as one mechanism stimulating societal change (e.g., Kennett et al. 2011;
Kohler and van der Leeuw 2007; Winterhalder et al. 2010). The growth of urban centers
also presents a complex ecological problem (Grimm et al. 2000; Zeder 1991); both
52
reducing agricultural activity in the urban core and expanding it in the periphery. The
degree that landscapes are altered is an empirical question heavily dependent upon local
context, including geological substrate, vegetation cover, and topographic controls on
hydrology and geomorphic processes. The sensitivity of landscapes to changing
anthropogenic and environmental conditions can only be determined through applied
geoarchaeological work.
Anthropogenic alteration of the landscape has featured prominently in models of
the emergence, persistence and transformation of ancient Maya sociopolitical and
economic systems (Demarest et al. 2004; Demarest 2006; Webster 2002) and empirical
evidence indicates that deforestation and erosion occurred in several parts of the tropical
Maya lowlands starting as early as the Late Preclassic Period (Anselmetti et al. 2007;
Beach 1998; Beach et al. 2006; Brenner et al. 2002; Curtis et al. 1996, 1998; Dunning et
al. 2002; Islebe et al. 1996; Mueller et al. 2010). Paleoclimatologists have also identified
intervals of greater or lesser rainfall during the Late Holocene that would have altered
vegetation cover and promoted erosion (Haug et al. 2001, 2003; Hodell et al. 1995, 2001,
2005; Mueller et al. 2009; Stahle et al. 2011; Webster et al. 2007). Complex land use
histories in the Maya Lowlands described in the last two decades have shown that the
ancient Maya adapted to local conditions of soil fertility, seasonal drought, and social
organization to produce multiple land use strategies, and that generalizations about Maya
agricultural practices often fail at inter-regional scales (Beach et al. 2006, 2008; Dunning
et al. 2002; Fedick 1996a; Fedick and Ford 1990). Therefore, explaining the emergence
and disintegration of individual Maya polities requires site-specific geoarchaeological
records integrated with cultural histories and climate records.
53
In this chapter I explore landscape changes before, during and after the formation
of the Classic Period Maya center of Uxbenká. The cultural chronology framing this
discussion draws from Demarest (2004:13) but is modified to follow the Late Archaic
and Middle Preclassic Period divisions proposed by Lohse et al. (2006; see Table 2.1)
The urban core of Uxbenká consists of six plaza groups that were carved from ridgelines
in this hilly landscape (Figure 3.1). Group A contains the remnants of 23 carved
sandstone stela dating to the Early and Late Classic periods and is presumed to be the
main ceremonial locus at the site (Prufer et al. 2011). Groups B-F are a contiguous
arrangement of plazas running along a ridgeline roughly 400m to the northwest of Group
A. The Group B plaza is a flattened hilltop and is surrounded by a series of range
structures and a large platform mound at its northern extent. A ballcourt dominates the
southern extent of the plaza. A second ballcourt is evident in the Group D plaza.
Construction in Uxbenká’s urban core began in the Late Preclassic, with the earliest
known structure in Group A dating to 60 cal BC - cal AD 220 (Culleton et al. 2012). The
massive effort of leveling and expanding ridgelines to form the Group B and D plazas
occurred slightly later, but still at the end of the Late Preclassic at cal AD 60-310 and cal
AD 20-240, respectively. There was a flurry of replastering and plaza renovation activity
until the first part of the Early Classic Period, and then less evidence for building activity
between cal AD 350-550. Architectural modifications are documented at Groups A, B,
and D after AD 550, including extensive plastering of plaza floors, laying paving stones,
and the augmentation of facades on existing structures. The latest dedicatory date
preserved on stelae at Group A indicatesthat monument carving continued until AD 781.
Political disintegration and the abandonment of this city in the Terminal Classic are
54
Figure 3.1. The Uxbenká site core, showing locations of geoarchaeological excavations in A) the core,and B) the Cochil Bul area to the north (basemaps by C. Ebert).
topics of ongoing research at Uxbenká, but there is currently no evidence for a Post-
Classic (after AD 1000) occupation of the site. The work presented here provides a
broader context for interpreting the urban and agricultural ecology of this small Maya
center.
55
Climatic Context
The Late Holocene climate history of Mesoamerica has been rapidly developing
since the mid-1990s with increasing attempts to explain major cultural transformations
with climate events. This is particularly the case with a series of Terminal Classic
droughts and the sociopolitical disintegration of many Maya polities (e.g., Gill 2000;
Haug et al. 2001, 2003; Hodell et al. 1995, 2001, 2005; Webster et al. 2007). Cultural
adaptations to changing climatic conditions (e.g., agricultural intensification) may have a
large effect on the landscape and are known to influence landscape transformations
directly due to vegetation change (Mueller et al. 2009). The three records considered here
– the Cariaco Basin marine Ti record (Haug et al. 2001, 2003); the Lake Chichancanab,
Mexico, core sediment density record (Hodell et al. 2005); and the Macal Chasm, Belize,
speleothem record (specifically the luminescence proxy; Webster et al. 2007) – are the
most proximate to the site of Uxbenká and they cover the time span of interest (roughly
the last 3500 years). Each provides a slightly different proxy for precipitation. General
features of the three records are in fair agreement, but often specific details differ
between the records (e.g., the timing or structure of Terminal Classic droughts), which is
due to the combination of the differing sensitivity of each proxy to climate change,
varying chronological precision in the underlying age models, and the potential for
regional climate events to have locally distinct and possibly contradictory expressions.
The chronological resolution of the geomorphic record presented here is at the
multi-decadal to centennial scale given the pace of many soil-formation processes and the
reliance on AMS dates on charcoal within the paleosols to determine age. Each AMS 14C
date occurs within a span of soil-formation rather than the exact age of the soil.
56
Therefore, the annual and decadal features of the climate records (and the conflicts
between them at these scales) are de-emphasized in favor of the broader temporal
patterns that are potentially linked to changes in soil stability and instability.
After the generally warmer condition during the middle Holocene Thermal
Maximum, the Cariaco Basin Ti record indicates an increase in El Niño/Southern
Oscillation (ENSO) intensity and variability from ca. 3000 BC, with the highest ENSO
intensity between 1500 BC and 400 BC (Haug et al. 2001, 2003). This era of climate
vicissitudes spans the end of the Late Archaic and most of the Middle Preclassic periods,
and is also seen in the early sections of the Macal Chasm (MC) speleothem luminescence
record (starting from ca. 1200 BC; Webster et al. 2007) and the Lake Chichancanab (LC)
density record (starting from ca. 850 BC; Hodell et al. 2005). Two severe droughts in the
Late Archaic are inferred from Cariaco at 1200–1000 BC and 950–850 BC and may
correspond to the two drier periods in the MC record from 1200–1000 BC and 1000–800
BC. The Middle Preclassic Period appears to have experience a prolonged trend of
overall drying with marked wet-dry oscillations and punctuated drought episodes
between 700 and 500 BC indicated by Cariaco, and 800 and 600 BC in the MC
speleothem. Lake Chichancanab also reveals a series of dry episodes between 750 and
300 BC at roughly 100-yr intervals that bleed into the Late Preclassic period. During the
Late Preclassic period the three climate proxies show less obvious coherence, but it
appears that precipitation was variable during the centuries that opened and closed it
according to the LC and Cariaco cores. Drought events in the middle part of the Late
Preclassic are also suggested in Cariaco at 200–50 BC and in the MC speleothem at 50
BC–AD 150.
57
The Classic Period trends in the three records collectively suggest relatively
wetter conditions during the Early Classic (with a dry episode recorded on the MC
speleothem from AD 450–550 not seen in the other proxies) that give way to a general
drying trend persisting through the Late Classic. Here again the details of the records
conflict, but all record the driest period since the Middle Preclassic or the end of the Late
Preclassic (but of longer duration) from AD 700 to 850. Lake Chichancanab and the MC
speleothem both indicate extremely dry conditions into the Terminal Classic Period (AD
850–1000), though this period is punctuated by a relatively wet period in the Cariaco
record that corresponds to the Medieval Climatic Anomaly (Haug et al. 2001, 2003).
The History of Maya Land Use
Contemporary landscapes in the Maya region are the products of millennia of land
use decisions in the face of changing modes of agricultural production, demographic
pressure, local micro-environmental conditions, and climatic change (Beach et al. 2006;
Denevan 1992; Dunning 1996; Dunning and Beach 2000; Fedick 1996a; Fedick and Ford
1990; Wingard 1996). Forest clearance through the use of fire was and continues to be an
effective, labor-saving component of Maya subsistence systems (i.e., both in foraging and
food-production contexts; Nations 2006; Nations and Nigh 1980) and changing charcoal
abundance in lake and wetland cores indicate the intensity of forest burning throughout
the Holocene. Increased fire frequency in the Maya Lowlands at the beginning of the
Late Holocene (~2000 BC) correlates with pollen spectra showing increases in
domesticates (Zea sp., Manihot sp.), disturbance taxa (e.g., Graminaea, Cyperacea) and
declines in primary forest arboreal taxa (e.g., Moracaea, Urticacaea, Bursuraceae)
58
(Piperno and Pearsall 1998). Increasing soil erosion is indicated in several lake records
during this period in the Petén (Guatemala) and the Yucatán (Mexico) regions,
suggesting the emergence of long-fallow swidden agriculture in upland areas made
feasible by the drier Late Holocene climate (Piperno and Pearsall 1998; Rosenmeier et al.
2002a, 2002b).
By 1500 BC, regional adaptations to wetland agriculture became important,
notably in the bajos of northern Petén (Hansen 1993, 1994) and the lowland swamps of
northern Belize. Earlier research suggested extensive raised fields in the Pasión region of
Guatemala (Adams 1980; Adams et al. 1981) and at Pulltrouser Swamp in northern
Belize (Harrison 1993, 1996; Puleston 1978; Turner and Harrison 1983) dating primarily
to the Late Classic Period (AD 600–800). Further research suggests that many of these
are either natural landforms that were never cultivated, or in northern Belize were fields
drained by ditching in the Preclassic (~1000 BC), but were not raised per se in the
manner of chinampas (Dunning 1996; Dunning et al. 1991; Pohl and Bloom 1996; Pohl
et al. 1996; Pope et al. 1996). Drained fields on Albion Island, and Douglas, Cobweb, and
Pulltrouser swamps appear to have been completely inundated and abandoned by ~ 200
BC due to a rising water table (Pohl et al. 1996).
Landscape alteration accelerated in the Maya region after ~1000 BC, as
population pressure forced a shift to short-fallow agriculture, putting more land,
including less favorable hillslopes, under cultivation in some regions. Buried topsoils
dating to 1500 BC at La Milpa and Petexbatun indicate that soil instability and
sedimentation rates increased in response to agricultural intensification during the Middle
to Late Preclassic (1000 BC–AD 300; Beach et al. 2006; Dunning and Beach 2000;
59
Dunning et al. 1999). In the Petén lake records, inorganic sediment and charcoal
abundance due to the shift to short-fallow swidden is likely superimposed on the signal of
drier climate through the Late Holocene, demonstrating the complex linkages between
human alterations, vegetation cover, and geomorphic stability (Binford et al. 1987; Curtis
et al. 1998; Hodell et al. 1995, 2000; Rice 1993; Rosenmeier et al. 2002a, 2002b).
Behavioral responses to environmental degradation during the Preclassic to Early Classic
involved decentralization or out-migration to other regions, but soil retention structures
(e.g., terraces, check dams) do not appear to have been employed during this period
(Dunning and Beach 2000).
New polities were established during the Classic Period (AD 250-800) and
agricultural practices intensified from long- to short-fallow systems, amidst a backdrop of
growing population and increasingly dry and erratic climate from ~AD 1–1000 (Haug et
al. 2003; Hodell et al. 1995, 2000). In this context, diverse human responses to demands
on the land are evident and illustrate the complexity of Classic Maya political
disintegration. In the Copán Valley, cultivation spread from the productive “pockets” of
the valley floor, and eventually onto the hillslopes under steady demographic expansion,
overtaxing productive capacity and undermining the geomorphic stability of the soils
(Abrams and Rue 1988; Webster et al. 2000; Wingard 1996). Prolonged drought episodes
during the Late and Terminal Classic (AD 600–1000) further decreased vegetative cover
and exacerbated anthropogenic erosion, culminating in landslides that buried parts of the
Main Group under as much as 2 m of colluvium (Abrams and Rue 1988; Fash and Sharer
2003; Webster et al. 2000; Wingard 1996). In the Petén and Yucatán, lake cores show a
similar mass-wasting event represented by the “Maya clay” (Binford et al. 1987; Deevey
60
et al. 1979; Hodell et al. 1995, 2000), and in northern Belize the Preclassic drained fields
are capped by an analogous stratum (Pohl and Bloom 1996; Pohl et al. 1996; Pope et al.
1996). Centers in the vicinity of Petexbatun, in contrast, show no evidence of increased
erosion during this period despite intensive cropping and continual forest suppression
seen in pollen records (Beach et al. 2006; Demarest 2006; Dunning 1996; Dunning and
Beach 2000; Dunning et al. 1998). A sophisticated array of conservation measures
including terraces, check dams, and reservoirs at Petexbatun, La Milpa, and Tamarindito
allowed for sustained intensive agriculture without runaway environmental degradation.
The elaborately terraced landscapes around Caracól are another example of land
conservation in the face of intensive cultivation (Chase and Chase 1998; Chase et al.
2011; Healy et al. 1983).
In sum, multiple land use strategies, conservative and otherwise, were employed
until the Terminal Classic (AD 800–1000) in response to changing climate, local soil
characteristics, available technology and social organization, along with the perceived
need or desire to mitigate the effects of anthropogenic landscape alteration. Given the
array of local factors informing these decisions, we may expect that extrapolations from
one region’s landscape history to another’s will be inadequate to explain the
sociopolitical evolution of any one polity (Beach et al. 2006, 2008; Dunning 1996;
Dunning and Beach 2000; Fedick 1996b, 1996c; Fedick and Ford 1990). The site-
specific, empirically grounded work described here explores human adaptive responses to
natural and anthropogenic environmental change at Uxbenká, and helps elucidate the
other social and ecological factors that contributed to societal transformation.
61
Field Methods
Geoarchaeological investigations were carried out from 2007 to 2009 at Uxbenká,
focusing primarily in the site core amid the main civic/ceremonial architecture groups,
Groups A, B and D (Figure 3.1; Culleton 2008, 2009, 2010). The main aim of these
excavations was to expose geomorphic profiles that would allow cultural features (e.g.,
architecture, middens, etc.) and paleosols to be identified and described. Where possible,
excavation units were taken to bedrock. This was motivated by a desire to identify the
most ancient paleosols at the site and to understand the local effect of the bedrock on
erosion, deposition and soil genesis. Excavations were conducted initially in natural
levels, and sediments screened through ¼-inch wire mesh where possible. Screening all
of the heavy clay loam sediment would have been prohibitively time-consuming, so
subsamples of sediment were screened to recover artifacts when paleosols and other
depositional surfaces were encountered. Artifacts were most commonly recovered by
excavators at the trowel’s or shovel’s edge rather than from screens. Profiles were
recorded and described according to Birkeland (1999).
Chronology
Radiocarbon samples to establish the ages of palaesols and cultural features were
recovered from profiles, features or recovered soil samples, in most cases selecting
individual twigs or single charcoal pieces to avoid problems of mixed age samples (Table
3.1). Specimens were pre-treated and combusted along with known-age standards (e.g.,
OX1 oxalic acid, Queets A wood, FIRI-H) using routine ABA techniques for organics at
the University of Oregon. Sample gas was submitted to UC Irvine Keck Carbon Cycle
62
AMS Facility for graphitization and AMS 14C measurements. Conventional ages are
δ13C-corrected using values measured on the AMS according to the conventions of
Stuiver and Polach (1977). Ages were calibrated with the IntCal09 atmospheric curve
(Reimer et al. 2009) using OxCal 3.01 (Bronk Ramsey 1995, 2001). Most charcoal
specimens were recovered from identified A horizons of those soils, and therefore
estimate points when the soil was stable and accumulating organic matter over some span
of decades or centuries. For a specific exposed paleosol, these dates represent the
minimum age (i.e, terminus post quem) of their burial. Because many of these dates fall
into discrete clusters, paleosols are correlated between units and modeled multiple
paleosol ages as phases using OxCal to estimate the beginning, end and span in calibrated
years. A chronology of geomorphic stability and instability within the Uxbenká site core
is established from those estimates.
Table 3.1. Calibrated AMS 14C dates from paleosols
UCIAMS # Provenience Conventional Age ( 14C BP)
2σ range cal BC/AD
Late Archaic 67230 SubOp 09-1, AG12, 300 cmbd, L3. 3555±20 1960-1770 BC 57040 SubOp 08-1. AG3, L.3, 290-300 cmbd. SS11. 3070±15 1410–1290 BC 56355 SubOp 08-3. AG9, W Wall. Top of Bosh Lu'um, 95-105 cmbd. 2955±20 1270–1080 BC 67953 SubOp 09-1, AG12, 185-190 cmbd, Top of L3. 2900±15 1190-1010 BC 68835 SubOp 09-1, AG13, 280-285 cmbd, Base of Exc. 2875±15 1130-1000 BC 68833 SubOp 09-1, AG12, 201 cmbd, L3. 2810±15 1010-915 BC 57039 SubOp 08-1. AG3, L.3, 290-299 cmbd. SS10. 2810±15 1010–915 BC
Middle Preclassic 68834 SubOp 09-1, AG13, 170-175 cmbd, gray wedge. 2500±15 770-540 BC 76156 SubOp 09-1, AG12 120-125 cmbd, Top of L2 2490±20 770-520 BC
Late Preclassic/Early Classic 56350* SubOp 08-1. AG3, 138cmS/63cmE, L.3, 167cmbd. 1950±15 AD 1-85 57038 SubOp 08-1. AG3, L.3, 200-210 cmbd. SS4. 1830±15 AD 130-240 56354 SubOp 08-1. AG6, L.4, Bosh lu'um, 192cmbd .RC8. 1780±15 AD 140-330 57037 SubOp 08-1. AG6, L.3, 160-170cmbd. 1730±15 AD 250-390 57041 SubOp 08-1. AG6, Fea. 2, 160 cmbd. SS12A. 1725±15 AD 250-390
Late Classic 36946 AG1, Buried Soil in Zone 2, 100-105 cmbd 1470±50 AD 430-660 56357 SubOp 08-3. AG11, 145cmN in E Wall, 277cmbd. RC15. 1455±15 AD 570-645 56356 SubOp 08-3. AG11, 235cmbd. RC12. 1455±15 AD 570-645
Terminal Classic 56352 SubOp 08-3. AG8, 212-219cmS/85-95cmE, L.2, 180cmbd. RC6. 1120±15 AD 885-975 56353 SubOp 08-3. AG8, 228-232cmS/70-80cmE, L.2, 198cmbd. RC7. 1115±15 AD 890-980 36947 AG1, Fill at bedrock, Zone 1. 1110±30 AD 870-1020
63
Results
Bedrock Geology and the Geoarchaeology of Uxbenká
Over the course of several years of archaeological survey and excavation at
Uxbenká, the local expression of bedrock geology has been found to dominate
geomorphic processes of erosion, deposition, hydrology, and soil formation, as well as
influencing the architecture of settlements and the site core. As described above, the
sedimentary Toledo Beds comprise a range of interbedded mudstones, sandstones and
limestones that are close to horizontal, typically not dipping by more than about 10-15° in
the site vicinity. The mudstone strata (locally called nib in Mopan Maya) break down
readily to form new soils when exposed to weathering, which contributes to the
“paradoxical” fertility of the soils around Uxbenká (Hartshorn et al. 1984:76-77). Nib is
also easily excavated without metal tools, and was used as construction fill in structures
of all sizes, which requires a careful eye to distinguish from in situ bedrock during
excavation. Sandstone strata are generally more durable than the nib, even where they are
erodible and not well indurated. Resistant sandstone strata overlying friable nib result in
flattened hilltops with steeply eroded hillsides in some areas, e.g., at SG 1 to the
northwest of the site core. In many cases sandstone outcrops eroding from hillsides
provided building material for house mounds and other domestic structures. In the case of
SG 25 and SG 28 to the east of the site core, an indurated sandstone member with
squared vertical joints gives the appearance of deliberate construction, but was simply
augmented with a few courses of additional masonry to create a more impressive
appearance (Figure 3.2A). Sandstone blocks are the primary construction material for
core architecture at Uxbenká, and the large flat sandstone slabs exposed in Santa Cruz
64
Figure 3.2. Characteristic exposures of the Toledo Beds in the Uxbenká vicinity: A) Sandstone outcrops at SG 25 taking the form of natural steps (note rock hammer for scale); and B) near-vertical joints in the nib (mudstone) forming a sheer face in a drainage to the east of the site core.
Creek and to the south at SG 35 were an ideal source for the numerous stelae carved and
erected at Group A during the Classic Period. The similar character of sandstones at Nim
Li Punit may have also contributed to the prevalence of a stela tradition there.
Near-vertical jointing in the nib and sandstone is very common in the Uxbenká
area, and likely reflects compressional stress from the tectonic activity associated with
the uplift of the Maya Mountains since the Cretaceous (cf. Hartshorn et al. 1984:12;
Figure II-2). At small scales (1–100cm) these joints contribute to the friability of the nib
and the ease with which sandstone slabs can be excavated from these outcrops. At larger
scales (1-10 m), the joints may be expressions of the faulting itself, and they mark the
landscape with narrow vertical chasms that in some places dictate the hydrology by
capturing streams, and in others dominate soil processes by creating deep sediment traps
(Figure 3.2B). Transects excavated along hillslopes in the site core and just to the east
demonstrate this process. Augering on the west slope of the Ha’il Chepa drainage near
65
SG 26 indicates a stepped pattern to the horizontally bedded nib, so that soil depth can
vary from 10 to 150cm or more even over short distances (Culleton 2010; Figure 3.3A).
In the Uxbenká site core between Group A and B the stepped bedrock is punctuated with
multiple sediment-filled chasms that create the initial impression of purposeful terracing
(Figure 3.3B). In the course of excavating these putative terraces their geological origin
became clear, and further observations in the site vicinity have so far revealed no firm
evidence for agricultural terracing at Uxbenká. Instead it appears the natural sediment
traps serve practically the same soil conservation function, as well as providing the
paleosol sequences that span the last 3500 years at Uxbenká that form the body of
geoarchaeological data presented here.
Excavations in the Site Core
Eleven excavation units were placed in the site core in 2007 and 2008, most were
1 m wide trenches ranging from 3-13 m long that taken together form a composite
hillslope profile spanning ~65 horizontal meters and ~16 m of elevation (Culleton 2008,
2009). For the sake of clarity I divide the slope into four zones that roughly correspond to
what were first assumed to be separate terrace platforms, Zone 1 being the lowest
elevation and Zone 4 being the furthest upslope (Figure 3.3B) Zone 1 comprises two ~5m
wide steps on the hillside excavated to bedrock with two parallel trenches (AG1 and
AG2). The trenches exposed irregular channels in the nib bedrock running with the strike
of the hillslope and ranging from 1-2m wide and 50-100cm deep. Between the channels
the soil depth ranged from 10-20cm. No clear paleosols were exposed in the units, but the
few artifacts recovered (non-diagnostic ceramic body sherds) were lying almost directly
66
Figure 3.3. Transects showing the marked jointing in the bedrock A) east of the Uxbenká site in the Ha’il Chepa drainage, and B) in the site core itself.
on the bedrock, suggesting the channels were relatively free of sediment when the
artifacts were deposited. Charcoal recovered from fill at the base of the upper channel in
AG1 dates to the Terminal Classic (UCIAMS-36947; 2σ: cal AD 870-1020) likely
representing infilling after the abandonment of Uxbenká.
Zone 2 was the focus of fairly intensive work (including units AG1-6 and AG8)
as the excavations revealed a series of natural and cultural strata, and sediments
extending to a depth of nearly 3.5 m below the present surface (Figure 3.4). A distinct 7-
10 m wide trough delineated on the downslope side by a nib bedrock outcrop rising about
50 cm above the soil surface markthe area. The outcrop represents the uppermost extent
67
Figure 3.4. Composite profile of AG3 and AG4 west walls, showing sandstone alignment (Feature 1).
of an almost vertical 4 m nib bedrock face that forms one margin of a large joint in the
bedrock.
A buried soil dating to the Late Classic Period was encountered in the section of
AG1 that extended in to Zone 2 (UCIAMS- 36946; 2σ: cal AD 430-660). Units AG3-6
were placed from roughly 10m to the west of AG1, being initially oriented perpendicular
to the nib outcrop. The number of units excavated was increased to investigate a variety
of buried soils and features in the section. Of particular note, a linear alignment of
unworked sandstone slabs (Feature 1) was uncovered at 150-200 cmbd running
perpendicular to the nib wall and that continued without apparent breaks or corners for
more than 7 m. The composite profile shows the feature comprising 3-4 courses of
unmodified sandstone slabs (with a few limestone pieces) following the sloping surface
of the paleosol. This feature was built upon a dark, organic rich paleosol, and was
covered by a complex series of broken nib fill layers and less distinct but readily
68
identifiable strata. Overall the feature appeared to be intact, though slabs adjacent to the
nib wall were less coherent, suggesting a disturbance such as an earthquake and
associated hill slope slide. As exposed in AG6, Feature 1 was confirmed to be a linear
feature, and stratigraphically above the paleosol. The series of strata observed above the
paleosol and Feature 1 were also present in AG6. A small lens of burned sediment and
charcoal (Feature 2) was found on the surface of the cultural stratum immediately above
the paleosol. Charcoal samples from Feature 2 and the stratum on which it was deposited
gave identical Early Classic dates (UCIAMS-57041 and UCIAMS-57037; 2σ: cal AD
250-390). Both are stratigraphically superior to the sandstone alignment so these dates
are the upper bracket on the age of its construction. Charcoal dates on the paleosol below
the alignment in AG3 (SS4, 200-210 cmbd; UCIAMS-57038; 2σ, cal AD 130-240) and
AG6 (RC8, 192cmbd; UCIAMS-57038; 2σ, cal AD 140-330) suggest it is a Late
Preclassic occupation surface/cultural deposit. Feature 1 is bracketed by Late Preclassic
and Early Classic strata and the date of its construction can be estimated using a sequence
model in OxCal. It appears to have been constructed at the end of the Late Preclassic, or
possibly the very beginning of the Early Classic Period (1σ: cal AD 230-290; 2σ: cal AD
200-340). This may be contemporary with the purposeful burial of 1st to 2nd century AD
cache in SG 20, on the ridgetop immediately above Zone 2, which was buried under a 1.3
m-tall mound of mixed nib and soil fill sometime after ca. 135 cal AD (Prufer et al. 2011:
213-214).
Excavations continued in the center of AG3 and before reaching nib bedrock at
404 cmbd, a second diffuse darker layer consistent with a buried soil surface was
identified ~300cmbd. This stratum also contained a few non-diagnostic ceramic sherds,
69
indicating human use of the area before the construction of Feature 1. Two radiocarbon
dates between 290-300 cmbd date this to the Late Archaic (SS11; UCIAMS-57040; 2σ:
1410–1290 cal BC; and SS10; UCIAMS-57039; 2σ, 1010–915 cal BC), and with the
span between the dates suggesting the buried soil was a stable surface for as much as 500
years before being buried. This sparse deposit represents some of the earliest evidence of
occupation at Uxbenká. Sediments below this exhibited strong mottling and ped faces
were well coated with clays, typical of a well-developed tropical vertisol.
AG8 (3x1m) was placed about 20 m west of the main Zone 2 excavations, where
the linear depression narrows and begins to conform with the topography of the hillside.
Excavation revealed a similar overall pattern of strata to the units in Zone 2, as well as a
portion of another sandstone alignment (Figure 3.5). The profile in the south half of the
unit reveals an original hillslope surface that dipped down abruptly to form a channel,
which filled in over time with successive cultural and natural sediments. At a depth of
roughly 70-90 cmbs (140-160 cmbd), a stratum of reworked nib colluvium was
encountered that capped a buried A horizon. At the base of the A horizon was a distinct
charcoal-rich layer in the south half of the unit. These charcoal pieces were large (1-2 cm
diameter) compared to other strata, and gave the impression of a short-fallow milpa that
had been chopped and burned before being buried. Two charcoal samples, from the A
horizon (110 cmbs; 180 cmbd; UCIAMS-56352) and from the charcoal layer (128 cmbs;
198 cmbd; UCIAMS-56353), produced essentially identical dates at cal AD 885-980
(2σ). These dates are consistent with the Terminal Classic, and fall in line with a date on
the fill in Zone 1.
70
Figure 3.5. AG8 west wall profile, showing sandstone alignment (Feature 1).
Two less distinct but clearly recognizable soil surfaces were observed at ~190
cmbs (260 cmbd) and ~220 cmbs (290 cmbd). The southwest corner of AG8 cut into a
portion of another sandstone alignment (Feature 1) that appears to have been constructed
on top of this lowest buried soil, being made from 2-3 courses of stone in the observable
section. As with the feature in the main part of Zone 2, none of the stone slabs showed
signs of modification, and all appeared to be the common tabular pieces used in
71
structures around Uxbenká. Several stones were tilted ~45° from horizontal, suggesting
the feature had collapsed into the apparent depression in which fill was accumulating
during the Classic period. There is no obvious relationship between the two sandstone
features in Zone 2 as their function is unclear, but it seems likely that the two sandstone
alignments are contemporary (i.e., dating to the Late Preclassic), and served the same
purpose, whether architectural, agricultural or otherwise.
Excavations on the flat tier of Zone 3 deployed 3 trenches. AG 7 (3x1m) was
placed 4 m uphill of AG8 at the head of the slope between Zone 2 and Zone 3 to
investigate the relationship between the two areas in terms of soil sequences and bedrock
morphology. No paleosols were identified in the unit, which appeared to contain a single
stratum with A (Ap), Box, and C horizons over bedrock. Given the slope, it is likely that
the soil formed in colluvium that was continually moving downslope, but without net
gain or loss. The underlying bedrock in the unit was virtually horizontal, which is
consistent with the observed bedding of outcrops around the Uxbenká, but also suggests
there must be a fairly sharp vertical drop-off to the bedrock between AG7 and AG8. If so,
Zone 2 may be flanked on the uphill side by a nib bedrock face similar to that exposed on
its downhill side. AG10 (8x1m) was placed 6 m upslope of AG9 on the head of the slope
between Zone 3 and Zone 4, which is a steep drop-off in the bedrock, covered with a thin
veneer of topsoil (i.e., 5-10 cm). The excavation involved clearing the thin soil and
decomposing nib to expose the bedrock. As elsewhere in the site core, the bedding was
nearly horizontal and the vertical joints were oriented along the strike of the hill slope.
AG9 (3x1m) was placed in the center of the large flat section of Zone 3, 6 m
upslope of AG7. Excavation exposed a well-developed paleosol, overlain by a layer of
72
loose nib debris, in a similar sequence to that observed in AG3 and AG6 above the Late
Classic strata and AG8 above the Terminal Classic stratum (Figure 3.6). An AMS 14C
date on charcoal from the buried 2Ab horizon (95-105 cmbd) of 1270-1080 cal BC (2σ;
UCIAMS-56355) places it in the Late Archaic. As noted in the discussion of the Late
Archaic paleosol in AG3, this date is bracketed by two other dates from that stratum,
suggesting that the two paleosols are correlated. In the case of Zone 2 it is clear that after
a period of relative geomorphic stability and soil development in the Late Archaic, Zone
2 accumulated sediments until the next evidence of occupation in the Early Classic. In
Zone 3 it is not clear when the Late Archaic paleosol was buried; a few ceramic sherds
were found in both the paleosol and the stratum above the nib debris, but were not
diagnostic. If the various nib debris layers in other units could be correlated, it is possible
that the paleosol in AG9 remained stable and available for occupation into the Classic
Period. The well-developed Box horizon in the upper stratum suggests a longer period of
soil development, starting before the Classic Period, leaving open the possibility that the
paleosol was buried after initial land clearing by Late Archaic farmers, although this
could have occurred later in the Middle or Late Preclassic period.
Zone 4 comprises a large fissure just below the ridgeline of the hill between
Groups A and B. Steep eroding nib walls rise 2-3 m on either side above the accumulated
sediment that forms its floor. The feature originates as a narrow (1 m-wide) step in the
hillside at its east end (where the uphill end of AG10 begins), and broadens to 5-7 m wide
at the location of AG11, roughly 30 m to the west. It continues more than 50 m, turning
to follow the changing aspect of the hillslope from roughly southwest to northwest.
73
Figure 3.6. AG9 west wall profile.
AG11 was placed to cut a 1x5 m cross-section of the sediments, under the
assumption that it was a larger version of the trough in Zone 2 and could possibly retain a
buried cultural sequence. No distinct buried soil horizons were observed in the profile,
though there were diffuse concentrations of ceramic sherds at ~200 cmbd and 210-215
cmbd, as well as more frequent but scattered charcoal at 235 and 275 cmbd. The soil
remained largely structureless and consistent in color and texture despite slight variations
in the amount of loose nib inclusions. Overall this is consistent with cumulic soil
development in a continuously aggrading sediment. Two dates indicate sediment
deposition was more rapid here than on the terraces below: charcoal samples from 235
cmbd (RC12; UCIAMS-56356) and 277 cmbd (RC15; UCIAMS-56357) yielded
identical calibrated ages of cal AD 570-645 (2σ). These two dates are Late Classic, and
74
are in good agreement with the date on the buried soil in AG1 in Zone 2. Given the
cumulic nature of the sediments here, these two AMS dates do not reflect organic inputs
associated with in situ soil development. They likely are derived from the erosion of a
Late Classic soil upslope of the chasm some time after ca. cal AD 645.
Excavations Northeast of the Site Core, Cochil Bul
Excavations were conducted to the northeast of the site core in 2009 to expand
geoarchaeological work further outside of the site core (Figure 3.1B). The site on the
north side of the ridge between Groups A and B was chosen for excavation after
reconnaissance survey revealed a relatively flat section at the base of a hillslope bordered
on the downhill margin by a linear bedrock protrusion reminiscent of that in Zone 2 in the
site core. Though Group B is visible in the distance from this location, there are no
known architectural features on the hills and ridge-tops immediately surrounding the
basin.
Two 3x1m units and an auger probe were excavated to identify whether a similar
series of buried soils was preserved in the natural sediment trap. A 4-inch dimaeter auger
probe went 4 m below the present surface, encountering possible paleosols at roughly
110–130 cmbs, 200–220 cmbs, and 335 cmbs based on coloration and the presence of
abandoned root channels. Bedrock was not reached by 4 mbs, indicating the potential for
very early buried sediments in this area. AG12 and AG13 were laid out perpendicular to
the natural rise and abutting it to the east, in the same manner as units AG1 and AG3 in
Zone 2. AG12 revealed at least three discernable paleosols at ~125 cmbd, 185 cmbd and
275 cmbd, with ceramic sherds and charcoal pieces commonly dispersed in relatively low
75
densities within them (Figure 3.7). The upper soil (from the present surface down to the
2Atb) contained very few artifacts and likely represents post-Classic/historic
sedimentation from the slopes to the northeast. AMS 14C dates on two individual charcoal
samples from the 3Atb horizon date it to the Late Archaic period (UCIAMS-67953, 185-
190 cmbd, 2σ: 1190-1010 cal BC; UCIAMS-68833, 201 cmbd, 2σ: 1005-910 cal BC),
and the span of dates suggests the soil surface was stable for a period of as much as 300
years. These two dates fall in line with the ages of paleosols in the site core excavated in
Zones 2 and 3, as well as the date at the base of the excavation in AG13. A date on
charcoal towards the base of the unit below the 4Aoxb comes much earlier in the Late
Archaic (300 cmbd, UCIAMS-67230; 2σ: 1960-1770 cal BC), though no cultural
materials were recovered from within this paleosol. As such, the earliest empirical
evidence for human presence at Uxbenká is found on paleosols dating a span from ca.
1200-900 cal BC in the later part of the Late Archaic period.
AG13 followed a similar course but revealed slightly more complex stratigraphy
than AG12, probably attributable to local variations in colluviation and drainage. Two
paleosols were encountered. The first clear paleosol (2Atb) undulated across the profile,
appearing to pile up against the bedrock wall to the south end of the unit. This suggests
the south half of the paleosol could be colluvium from the natural rise, or alternatively
part of the 2Atb was scoured before the present soil unit was deposited. The contact
between the second and third paleosols was expressed as two wedges of sediment, a
grey/brown wedge associated with 2Atb in the north half of the unit and a yellow wedge
in the south half. A single charcoal AMS 14C date places the 2Atb soil in the Middle
Preclassic, which is the first paleosol dating to that time at Uxbenká (170-175 cmbd;
76
Figure 3.7. Composite profiles of Cochil Bul excavations AG12, AG13, and Auger Probe
UCIAMS-68834; 2σ: 770-540 cal BC). The deepest paleosol (4Aoxb) had a clear
horizontal upper boundary and dispersed highly eroded ceramic sherds and charcoal. A
single AMS 14C date places this in the Late Archaic period (280-285 cmbd; UCIAMS-
68835; 2σ: 1130-1000 cal BC) and correlates well with dates on paleosols in the site core
and AG12.
77
Discussion
Six distinct paleosols were identified in the urban core of Uxbenká, the earliest
dating to the Late Archaic Period prior to evidence for human occupation at the site and
the latest dating to the Terminal Classic, a time when archaeological evidence for human
presence in the region is sparse. A Bayesian model that combines stratigraphic
information with AMS 14C dates from within these soils is provided in Table 3.2. Age
estimates and the projected duration of soil formation should be taken as a minimum
given the relatively small number of radiocarbon dates on each soil and the statistical
probabilities of having dated the first and last events within each soil. These were
modeled as phases in OxCal and using the boundary function partly minimizes this
problem (see discussion between Steier and Rom [2000] and Bronk Ramsey [2000]). The
timing of the geomorphic changes is considered with respect to cultural and climatic
records in the following discussion (Culleton et al. 2012; Haug et al. 2001, 2003; Hodell
et al. 2005; Prufer et al. 2011; Webster et al. 2007) (Figure 3.8).
Table 3.2. Geochronology of paleosols in the Uxbenká site core.
Paleosol Exposures
Earliest Formation (2σ)
Latest Formation/ Burial (2 σ)
Dated Span Range (cal yr, 2 σ)
Mean (cal yr)
Late Archaic AG3, AG9, AG12, AG13 1720-1280 BC 970-620 BC 320-470 390 Middle Preclassic AG 12, AG13 970-620 BC 750-300 BC -10-170 60 Late Preclassic AG3, AG4, AG6 AD 10-240 AD 160-320 -5-95 30 Early Classic AG3, AG4, AG6 AD 210-360 AD 280-610 -5-80 25 Late Classic AG1, AG11 AD 280-610 AD 610-960 -10-180 50 Terminal Classic AG1, AG8 AD 610-960 AD 890-1160 -5-80 35
The earliest paleosol in this series dates to the Late Archaic Period and is
represented by deeply buried A horizons exposed in the site core between Groups A and
B in units AG3 and AG9, and the Cochil Bul in units AG12 and AG13. The initial
formation of this paleosol is poorly constrained between 1720-1280 cal BC, but occurs
78
Figure 3.8. Geomorphic stability and instability at Uxbenká compared to cultural chronology and climate records. Periods of soil formation and stability as estimated from AMS 14C dates are shown as grey blocks. A) Timing of major construction activities in the site core (2σ; from Culleton et al. 2012). B) Estimated onset of erosion events. C) Lake Chichancanab Core Density Record (Hodell et al. 2005). D) Macal Chasm speleothem luminescence record (Webster et al. 2007). E) Cariaco Basin core titanium record (Haug et al. 2001, 2003).
within the Late Archaic Period. There is no evidence for human activity in the
surrounding area at this time and the absence of cultural material in the soil is consistent
with this observation. Non-diagnostic ceramic sherds recovered from the upper portions
79
of this soil provide the earliest evidence for human occupation within the confines of the
Uxbenká site core. These sherds were most likely deposited in or on a natural soil surface
and I conservatively estimate their age to be near the end of deposition (970 to 620 BC).
This age represents the onset of A horizon burial. The estimated age of these ceramics
falls into the early part of the Middle Preclassic Period and corresponds well with the
earliest pottery found elsewhere in Belize and the Maya region more generally. Swasey,
Bolay, and Cunil ceramic traditions in northern Belize are found no earlier than ca. 1000
cal BC, and sometimes appear as late as 800 cal BC (Awe 1992; Clark and Cheetham
2002; Hammond et al. 1991; Lohse 2010; Rosenswig and Kennett 2008). Lohse (2010)
has noted that many of the less-securely dated contexts for early pottery occur in Late
Archaic age paleosols at the base of Middle Preclassic Period excavations. Age
determinations for this early pottery are often from charcoal picked from sediments
directly overlying bedrock. This suggests that many of the initial ceramic components
were deposited onto older surfaces and eventually mixed into the soils by natural and
cultural processes.
Soil formation during the Late Archaic (~1700 - 900 BC) corresponds to moist
conditions evident in the Cariaco Basin and Macal Chasm records (Haug et al. 2001,
2003; Webster et al. 2007) and this would have promoted vegetation coverage and
inhibited erosion. There is also no evidence for human occupation or land use in this area
until the first appearance of pottery evident in the upper portions of this deposit dating to
~900-800 BC. The coincident appearance of pottery and erosion may signal the
appearance of pioneering Maya groups moving into the area and destabilizing the
landscape. Deforestation, landscape destabilization, erosion, and increased sediment load
80
in river systems is associated with the initial colonization of farmers elsewhere in
Mesoamerica (Joyce and Mueller 1992; Kennett et al. 2010; Neff et al. 2006) and also in
the Maya lowlands specifically (Jones 1994; Pohl et al. 1996; Pope et al. 1996, 2001).
The burial of the Late Archaic soil at Uxbenká also coincides with droughts evident in
both the Cariaco and Macal Chasm records at the end of the Late Archaic Period and the
earliest Middle Preclassic Period and this would have exacerbated any anthropogenic
impacts at this time. Although the effects of forest clearing by extensive and possibly
mobile farming communities on the erosional regime in this area is difficult to estimate,
the coincident appearance of pottery and increased erosion is highly suggestive.
However, one cannot rule out the possibility that the destabilization of the landscape at
this time was driven largely by drought. This is a topic for future work.
The Middle Preclassic paleosol is currently known from units AG12 and AG13 in
the Cochil Bul area to the north of the site core, and so may be a fairly localized
phenomenon. The burial of the Late Archaic paleosol between 970 and 620 cal BC marks
the earliest possible timing for the beginning of Middle Preclassic soil formation, and,
though only constrained by two AMS 14C dates, would have been a stable surface until
750-300 cal BC. As with the Late Archaic paleosol, ceramic sherds and chert debitage
indicate a cultural component in the vicinity of the later Uxbenká urban core in the
Middle Preclassic (at least by ~300 cal BC) that predates the earliest architectural
sequences at the site between ~60 BC and AD 220 (Culleton et al. 2012). Climatic
conditions through this period were quite variable, with several dry episodes
superimposed on a broader drying trend. The 2σ range of the two AMS dates from the
paleosol, 770-520 cal BC, corresponds with two severe droughts in the Lake
81
Chichancanab record and the pronounced decline in precipitation in the Macal Chasm
speleothem. Given the range of potential ages for the burial of this soil and variability in
the climate records in the Middle Preclassic, it is possible that after a brief period of soil
development it was buried by anthropogenic or drought-induced erosion at ~300 cal BC.
The series of natural and cultural strata exposed in AG3, AG4 and AG 6 in the
site core are contemporary with the establishment and expansion of the urban core at
Uxbenká. The estimate for the beginning of soil formation in the Late Preclassic paleosol
is cal AD 10-240, by which time clearing and leveling activities had already taken place
on the hilltop sites of Groups A, B, and D (Culleton et al. 2012). In addition to the
accumulated cultural materials in the A horizon, it is overlain by the 3-4 course alignment
of sandstone slabs (Feature 1), which is modeled to have been constructed between cal
AD 180 and 340 at the end of the Late Preclassic. The deposition and occupation of an
Early Classic soil and burn feature post-dates Feature 1, and initial deposition is
estimated at cal AD 210-360. This soil represents the surface on which residents of
Uxbenká carried out their daily activities during the Early Classic, as indicated by the
presence of broken pottery, charcoal, and a relatively slower accumulation of hillslope
colluvium during this time.
Based on the available data the landscape during the Late Preclassic and Classic
periods was relatively stable. Agricultural systems were well established in the Maya
region by the Late Preclassic and Early Classic Periods and building activities at
Uxbenká suggest a thriving population that was generating enough surplus to maintain
building campaigns directed by the ruling elite at this location (Culleton et al. 2012;
Prufer et al. 2011). At much larger Maya sites, active building programs are associated
82
with clear indications of intensified agriculture and soil conservation strategies in areas
peripheral to the urban core (e.g., terracing at Caracol and Petexbatun; Chase and Chase
1998; Chase et al. 2011; Dunning et al. 2002; Healy et al. 1983). Conservation
mechanisms do not appear to have been put in place at Uxbenká, and at face value one
can assume that the natural sediment traps in the site core obviated the need for
constructed terraces. However, the Early Classic paleosol is fairly thin and doesn’t appear
to represent a stratum of continually aggrading slope-wash colluvium. Although the
chasm in Zone 2 does act as soil retention feature there doesn’t appear to be evidence for
exceptional rates of erosion and deposition there during the Early Classic. I suspect this
was due to incompatibilities between civic-ceremonial activities and swidden agriculture.
During the Early Classic as the Uxbenká site core developed into an urbanized landscape,
regular clearing and burning for annual crops was likely relegated to more peripheral
locations and closer to domestic compounds positioned on hilltops outside the city center.
It seems unlikely that the urban core was allowed to return to high forest during the
Classic because maintaining an open viewshed within the civic-ceremonial core must
have been a priority for the ruling elites. A possible alternative would be a form of
arboriculture that kept economically or ritually important species (e.g., cacao, avocado,
mango, ramon) within the site core for the benefit of the elites, perhaps a version of the
“forest garden” originally proposed by Puleston (1978, 1982) and more recently
promoted by Ford (2005) and others (e.g., Fedick 1996c; Wyatt 2008; for persistent
effects of ancient forestry practices see also Ross 2011; Ross and Rangel 2011). This
change in land use led to local soil stabilization and decreased erosion even as
agricultural production in the larger polity intensified during the Classic Period.
83
The thick wedge of mixed dark soil and bedrock debris that covers the stable
Early Classic soil in Zone 2 of the site core (exposed in AG3 and AG4) suggests an
episode of mass-wasting and colluviation during the Classic Period estimated to at cal
AD 280-610. The colluvial deposit is generally consistent with the soil and bedrock
response to forest clearing described here, where topsoil runs off and the bedrock rapidly
breaks down to form new soil, but this is inconsistent with the urban setting by the Early
Classic. While the dry episodes recorded in the Macal Chasm speleothem could have
contributed to an erosional event toward the end of the Early Classic, evidence for a
possible tectonic event (exposed in AG3 and AG4) is a more compelling trigger for
landscape destabilization in an otherwise stable setting. Taken together, the disturbance
to the sandstone alignment, the discontinuity in the deeper sediments and the colluvial
stratum suggest an earthquake caused the bedrock to shift and the sediments to slump at
some point during or after the Early Classic Period sometime between cal AD 280 and
610. The potential for rapid nib mass-wasting was witnessed locally after a magnitude 7.3
earthquake struck on May 28, 2009, with its epicenter off the coast of Honduras. During a
survey in east of the site core the Ha’il Ayin drainage (cf. Figure 3.2B) two weeks later I
observed multiple scree piles and displaced boulders representing hundreds of cubic
meters of debris in the stream channel. It seems likely that the population of Uxbenká
witnessed a similar event in the site core during the Early Classic.
The Late Classic A horizon exposed in AG1 formed in a parent material with
fewer clasts than the colluvium that buried the Early Classic soil after cal AD 280-610,
suggesting it was formed from gradually accumulating slopewash rather than mass-
wasting. Two charcoal dates in a deeply buried section exposed in AG11 indicate the
84
presence of a Late Classic soil that was the source of that material on the slope above it.
The three dates combined into a phase estimate the end of soil formation during the Late
Classic or into the Terminal Classic at cal AD 610-960, after which time it was buried. It
was a time of increasing dryness evident in the Cariaco, Lake Chichancanab and Macal
Chasm climate records and the interval was punctuated by a series of marked droughts.
The chronology of construction activities in the Uxbenká site core is not well known for
the Late Classic Period, but stone monuments continued to be carved and dedicated until
AD 781, and architecture and artifact assemblages evident on the surface of the site
indicate that it still remained an urban space devoted to civic and ceremonial functions as
well as maintenance of prestige tree-crops for elite use. Slight increases in slopewash
associated with the Late Classic soil probably reflect the combination of increased aridity
seen in the climate records and erosion/deposition processes establishing a new
equilibrium after the possible Early Classic tectonic event.
Terminal Classic deposits in the site core comprise a buried charcoal-rich layer in
AG8, and a date representing in-filling of bedrock channels in Zone 1 of the site core.
The stratum exposed in AG8 is formed on a sediment deposit that buried a sandstone
alignment now partly exposed in the corner of the unit. Though similar to the feature
exposed in AG3, AG4, and AG6, it’s impossible to say whether they are
contemporaneous (therefore dating to the Late Preclassic), or shared the same obscure
function. The Terminal Classic stratum is characterized by a concentration of charred 1-2
cm diameter sticks that bears a strong resemblance to a milpa that was then rapidly
buried. The fill in the lower section of AG1 appears to have been deposited during the
Terminal Classic, possibly from slopewash derived from the uphill depression. The date
85
range constraining initial deposition is cal AD 610-960, and the burial of the deposit by
later colluvium is estimated to have begun from cal AD 890 to 1160. The Terminal
Classic is a period of extremely dry conditions, and it is likely that this contributed to
reduced vegetation cover around Uxbenká, making the soils more prone to erosion.
However, it’s unclear whether there was still a substantial population in the region by this
point, as there is no clear evidence of occupation in the site core at this time and stone
monument production had come to a halt. It is possible, and perhaps even suggested by
the presence of the burned layer itself, that smaller groups of farmers still resided in the
area during the Terminal Classic after the civic and ceremonial core was abandoned. The
Terminal Classic A horizon and subsequent sediment wasting may then be the result of
clearing and burning during a period of already prolonged drought that magnified the
effects of swidden agriculture at that time.
Conclusions
The geoarchaeological work at Uxbenká has defined two episodes of cultural
activity that precede the earliest evidence for the leveling and construction of buildings in
the urban core. Non-diagnostic ceramic sherds recovered from these A horizons provide
the earliest evidence for human occupation in what later became the urban center. This is
currently the earliest evidence for human activity in the area and is consistent with the
hypothesis that a small farming population first colonized the area between ~900 and 800
BC. This pioneering agricultural activity also occurred during a dry climatic interval that
may have destabilized the landscape further. Soil stability during the Middle Preclassic
(~770-520 cal BC) occurred during a drying trend that was punctuated by several severe
86
dry periods. This suggests that the landscape is fairly resilient under naturally dry
conditions. Destabilization again coincided with the appearance of pottery and stone tools
in the sediments at ~300 cal BC, but also with one of the more severe drying trends that
likely contributed to deforestation and erosion. I argue that the absence of agricultural
terraces and other soil retention features in the area surrounding the urban core results
from naturally occurring soil retention features and the rapid decomposition of the
mudstone bedrock favoring soil replenishment. I further argue that the overall stability of
the landscape in the urban core between ~60 BC and AD 900 resulted from the absence
or reduction of swidden cultivation in what was essentially an urbanized landscape used
for civic-ceremonial activities and possibly stabilized by urban gardens and the
cultivation of economically valuable tree crops. An episode of mass-wasting in the urban
core occurred during the Early Classic sometime between cal AD 280 and 610, and is
attributed to tectonic activity and associated hillslope failure, rather than human activities
in the site core. Increased erosion and the burial of the Late Classic Period landscape is
coincident with increasing evidence for swidden agriculture in the site core, possibly by a
remnant or returning population of farmers after the political collapse of Uxbenká that
occurred in the context of climatic and social instability during the Terminal Classic
Period.
87
CHAPTER IV
MAIZE AGROECOLOGY AND POPULATION ESTIMATES FOR THE ANCIENT
MAYA POLITY OF UXBENKÁ, BELIZE
This chapter was prepared as an unpublished co-authored manuscript with Dr.
Bruce Winterhalder, Claire Ebert, Dr. Prufer, and Dr. Kennett. I conducted the field work
to select maize plots and to collect soil samples, quantify harvest yields, and to organize
and analyze the yield and soil chemistry data. I also conducted the analyses to convert
estimated maize yields into the population estimates and predictions of settlement density
in the project area. Dr. Winterhalder contributed to the field research design and sampling
strategy, and provided guidance on integrating aspects of demographic theory in
anthropology with the maize population estimates. Claire Ebert organized the yield data
in a GIS database to produce the yield rasters, landscape coverages, and summary yield
calculations. Dr. Prufer oversaw field work, and Dr. Kennett provided insights into the
broader application of popualtion estimates in archaeological contexts.
Any explanatory model for the development and decline of human societies must
come to terms with the Malthusian problem of food limitations (Wood 1998). Whether or
not models are explicitly embedded within a neo-Darwinian evolutionary framework, the
role of changing population size and density are important variables in key developments
in human prehistory (e.g., Dumond 1975; Johnson and Earle 1987; Turchin 2003; Weiss
1976). These include the expansion of anatomically modern humans across the globe and
their concomitant ecological consequences (Burney and Flannery 2005; Erlandson 2001;
Fitzhugh and Kennett 2010; Goebel et al. 2008; Kennett et al. 2006; Kirch 2000; Martin
88
2005; Steele 2010), the transition to agriculture and subsequent spread of agricultural
populations from multiple centers (Barker 2008; Bellwood 2005; Childe 1928, 1951;
Diamond and Bellwood 2006; Kennett and Winterhalder 2006; Piperno and Pearsall
1998; Smith 1998, 2001; Trigger 2003; Zeder 1991), intensification of food production
and the emergence of social and technological complexity (Arnold 1992; Boserup 1965;
Carniero 1970; Cohen 1977; Kennett 2005), and the integration and decline of
institutions and state-level societies (Demarest 2004, 2007; Johnson and Earle 1987;
Kennett and Kennett 2006).
Population change has been cited on a conceptual level as either a cause or a
consequence of sociopolitical change, suggesting at a minimum a dynamic relationship
between population density and sociopolitical formations (Turchin 2003). Models of
Human Behavioral Ecology (HBE), including the Ideal Free Distribution (IFD) and the
Ideal Despotic Distribution (IDD), formalize explicit relationships between population
density and access to suitably productive habitats, and population dependent decreases in
habitat suitability (e.g., Kennett and Winterhalder 2008; Kennett et al. 2009; McClure et
al. 2006; Sutherland 1996; Winterhalder et al. 2010). Advances in theoretical population
biology and computational modeling allow for the exploration of long-term interactions
of ecological, demographic and social variables in past societies and the dynamic effects
on human decision making (Lee et al. 2008, 2009; Puleston and Tuljapurkar 2008;
Tuljapurkar et al. 2007). Meaningful applications of such models and simulations to a
specific prehistoric context must be guided by empirical data that archeologists and
human ecologists can provide about past environmental conditions, technological
organization, land-use patterns, settlement structure and population.
89
The consideration of population dynamics in ancient Maya society is strongly
dependent upon the ecological constraints of maize agriculture in the Neotropics and its
relationship to the emergence of social inequality and complex political systems. The
early 20th century notion of Maya polities as vacant ceremonial centers was based on the
assumption that the extensive swidden system of maize production could not support
substantial populations (Culbert and Rice 1990: xix). However, settlement surveys in the
1940s and 1950s brought to light large numbers of house mounds that indicated greater
populations than previously thought, and by the early 1980s evidence of agricultural
intensification in the form of raised fields and constructed terraces suggested the potential
for higher levels of food production than had been assumed from ethnohistoric accounts
(Adams 1980; Adams et al. 1981; Chase and Chase 1998; Harrison 1993, 1996; Puleston
1978; Turner and Harrison 1983). This evidence suggested that population pressure –
often represented theoretically by Bosreup’s (1965) model of intensified food production
– was a prime mover in the emergence of Maya sociopolitical complexity. State-level
development started sometime in the Preclassic Period, when the demands for centralized
labor and resource management provided conditions for political hierarchies to develop
(e.g., Adams 1977; Demarest 2007:162; Turner and Harrison 1978).
The role of population pressure in both the emergence and decline of ancient
Maya polities (and other state-level societies) is contentious for theoretical and empirical
reasons. Cowgill (1975a, b) argued that the assumption of inevitable population growth
isn’t borne out in ethnographically known small agricultural groups thought to be
comparable to Preclassic Mesoamerican peoples. Rather, most small farming populations
effectively maintain growth rates below a potential maximum through a number of
90
biological and social mechanisms. The conditional nature of that rate is seen when
populations expand rapidly in new environments or new sociopolitical contexts, but
under most circumstances constant population growth cannot be assumed, and is argubaly
a “non-explanation” of social change (Cowgill 1975a). At the other end of the arc of state
development and collapse, Webster (1985) criticizes the idea that the economic burden of
non-productive elites and specialists triggered sociopolitical tensions and the Maya
collapse. The problem is defining the population size and demand for resources with
respect to the productive capacity of the land and the agricultural system in place during
the Classic Period. Characterizing the relationships between these variables is key to
understanding the dynamic interaction between population density and sociopolitical
change.
Archaeologists have estimated ancient population sizes from various lines of
evidence including skeletal remains (cf. Wood et al. 1992), frequency of radiocarbon
dated components (Erlandson et al. 2001; Rick 1987), summed radiocarbon probabilities
(Buchanan et al. 2008; Shennan and Edinborough 2007), artifact consumption patterns,
and projecting ethnographic population estimates into the past (Haviland 1969). In the
Maya region settlement surveys at Tikal in the 1960s and 1970s (e.g., Haviland 1969,
1972), were used to derive population estimates from the number of domestic structures
occupied through time, making assumptions about the relationship between structure size
and function, number of occupants, and finally extrapolating from sampled portions of
the landscape to the entire populated area of the polity (e.g., Culbert and Rice 1990, and
studies therein). The settlement method of population estimation has the advantage of
directly reflecting past human presence on the landscape. The connection between
91
domestic structure density and population density across a region makes intuitive sense,
but it does require a robust survey coverage that ideally employs test-pitting and artifact
recovery to establish an occupational chronology (Culbert and Rice 1990; Webster et al.
2000).
Practical challenges to achieving such a settlement coverage in the Maya
lowlands are many and include: 1) poor visibility of low house platforms in secondary
tropical forest or scrubby bush; 2) the potential for buried components (e.g., Ashmore et
al. 1990 for Quiriguá); 3) absence of datable materials or ambivalence towards
chronological methods (e.g. with respect to obsidian hydration (see Braswell [1992,
1996] vs. Webster et al. [2004]); and 4) the large investment of time and financial
resources necessary (Webster et al. 2000). Arguably, few ancient Maya centers have
received the years of focused investigation that would be required to produce an
“adequate” settlement sample, perhaps with the exception of two of the largest, Tikal and
Copán (see Webster et al. 2000)
Estimating population size through study of the agricultural potential of a Maya
polity’s resource catchment provides another route of inquiry when settlement data is
lacking. Importantly, it can serve as an independent line of evidence to test
archaeologically derived estimates. One method is to iteratively model population change
based on agricultural productivity, informed by in-field soil survey, estimations of
erosion and recovery rates, mode of production (including fallow time and level of
intensification) and other ecological variables (Kohler and van der Leeuw 2007; Wingard
1992, 1995). Simulations allow for testing the effects of individual variables to isolate
and identify the most influential factors in a complex system. They are necessarily
92
diachronic and therefore amenable to historically, processually and evolutionarily-
oriented archaeological investigations. They often lead to unexpected insights into the
relationships between complex sets of processes (Kennett and Winterhalder 2006; Lee et
al. 2008, 2009; Puleston and Tuljapurkar 2008; Tuljapurkar et al. 2007; Webster et al.
2000; Winterhalder et al.1988; Winterhalder and Goland 1993; Winterhalder and Lu
1997).
Various authors have worked from ethnographic and ethnohistoric data on maize
(or other crop) production to estimate carrying capacity of a presumed area of cultivated
land around ancient Maya polities (e.g., Cancian 1965; Carter 1996; Puleston 1982;
Reina 1967; Reina and Hill 1980; Stadelman 1960; Tax 1954). Applying such data to a
specific archaeological setting requires careful evaluation of the comparability of the
ecological zones being considered (climate, geology, soils) and the mode of production
being practiced (e.g., level of mechanization, land tenure system, market engagement).
The site of Uxbenká provides a unique situation where contemporary Maya farmers from
the village of Santa Cruz are cultivating maize on the same lands as their ancient
counterparts in a largely non-mechanized swidden subsistence system. Here I use data on
maize yields collected in 2009 and 2010 to estimate overall yields for the lands around
Uxbenká and the maximal population density the ancient polity could have supported at
its height during the Late Classic period (AD 600-800). This effort to derive a synchronic
estimate of maximal possible population at Uxbenká is not considered to be a definitive
statement, but the first step towards developing more complex and demographically-
informed population models in the future (e.g., Lee et al. 2008, 2009; Puleston and
Tuljapurkar 2008; Tuljapurkar et al. 2007). Analysis of the local-level factors (soil
93
characteristics, slope, planting technique, etc.) that affect agricultural yields is an equally
important objective of this research. Better understanding of the causes of yield variation
is also an important and generalizable result of empirical work on ancient and
contemporary Maya food-production systems.
Setting and Background
The Maya village of Santa Cruz is located in Toledo District in southern Belize
(Figure 1.2). Approximately 400 people live in this village, which is located between the
neighboring communities of San Jose, Santa Elena and San Antonio. These reservation
lands have been cultivated under a traditional communal land tenure system for many
generations (Wainwright 2007). Government of Belize census data and fieldwork in 2006
by Wainwright (2007) indicate that the population is primarily Mopan (86%) and
K’ek’chi (14%) Maya. The typical seasonal round of maize cultivation in Santa Cruz
generally adheres to the patterns described by Wilk (1984, 1991) for other K’ek’chi and
Mopan Maya subsistence farmers in the Toledo District. It begins in the driest months of
the year (February/March averaging 40-70 mm/mo; Heyman and Kjerfve 1999; Wright et
al. 1959) when community members decide through informal discussions where to clear
land for their milpas (typically ranging from ~1-1.5 ha). Individuals or labor-exchange
groups cut patches of secondary forest or high bush. Land is cleared primarily by hand
with machetes, however in recent years a few (i.e., <5) chainsaws have been purchased
by individuals and are sometimes used to fell larger trees. Ideally the felled vegetation is
burned a week or two before the onset of the rainy season in May or June, and fields are
planted shortly before rains are expected to begin. This crop grows through the wet
94
summer months when rainfall ranges from 400 to 700 mm/mo (Hartshorn et al. 1984;
Heyman and Kjerfve 1999; Wright et al. 1959) and cobs are dried on the stalk to be
harvested starting in late September or early October.
At this time a second crop is planted in matahambre, which is similar to milpa,
but the felled vegetation is left as a mulch rather than burned. The matahambre maize
crop grows slower than the milpa crop due to the cooler and drier weather and it is
available to harvest in February. In the ethnographic literature the term matahambre
typically refers to a second maize crop planted on seasonally inundated floodplains and
levees (e.g., Reina 1967; Wilk 1991). The wet season planting done in upland settings
around Santa Cruz and other villages does not fit the classical definition of matahambre
except in its literal Spanish sense of ‘killing hunger’.
Arable soils around Santa Cruz are derived from the Toledo Beds, a series of
Tertiary interbedded calcareous mudstones, sandstones and shales that are bordered to the
south by a prominent Cretaceous limestone karst ridge (Keller et al. 2003; Miller 1996;
Wright et al. 1959). The karst, locally known as “The Rock Patch,” contains several
caves that are the subject of on-going archaeological research (Prufer et al. 2011). This
karst ridge dominates the drainage of the largest local stream, Rio Blanco, which flows
with its tributaries over the Tertiary beds south until meeting the southwest-northeast
trending ridge where it abruptly turns to the east. Eventually the Rio Blanco enters the
karst at Oke’bal Ha Cave and exits as Blue Creek to the south at Hokeb Ha Cave (Miller
1996). The rock patch itself is generally considered too steep and the soils too thin for
cultivation by most Santa Cruz farmers, though it is used as a source of forest products
95
for house construction, traditional medicinal plants, and small game hunting (TMCC
1992; cf. Steinberg 1998 for similar forest use in San José village to the north).
No formal ethnopedological study has been conducted in Santa Cruz, but it is
known that farmers make many fine and broad distinctions among the arable soils around
the village. The broadest practical distinction is between the box lu’um, well-drained
black clay loams largely distributed to the north of the village and at the base of the rock
patch, and the chik lu’um, poorly-drained oxidized reds soils primarily found in the
village itself and to the south within ~500-750 m of Rio Blanco. Box lu’um is favored for
almost any crop, whereas the heavy chik lu’um is primarily devoted to dry rice crops and
rarely for maize, which produces poorly under waterlogged conditions.
The seed stock for the milpa and matahambre crops are either saved by farmers
from previous harvests, traded or purchased within the village or less often acquired from
farther afield (e.g., from cobañeros in Guatemala). Mopan Maya farmers refer to some of
these varieties as “hybrid” corn, but it is unclear if these represent industrially-bred lines,
and more importantly whether they would remain true to type after multiple years of
cultivation. A great deal of ethnographic work would be required to identify the many
distinct land-races in circulation in Santa Cruz village, but they can be broadly
characterized as either long (e.g., shanil nul, box holoch) or short varieties (e.g.,
chaparro, bejuco). These names simultaneously refer to both the time to harvest and the
length of the husk with respect to the length of the cob. No significant differences in
yields have been observed between them. The decision to plant one or the other appears
to depend mainly on the farmer’s estimate of when the extant household supply of dried
maize will run out. Harvesting a short corn 2-3 weeks earlier may be the difference
96
between sustenance and temporary shortfall. The timing of harvest can also be disrupted
by delays in planting caused by late rains, scheduling conflicts with wage labor
commitments outside the village, illness, and other factors. Selecting a short variety helps
mitigate the late planting. A major disadvantage of short corn is greater susceptibility to
weevils and rot because the shorter husk provides less protection than in a long corn.
Unlike long varieties, short varieties cannot be saved for more than ~6 months and must
be replanted with each milpa and matahambre crop to save the seed, sometimes in a
small (e.g., 25 x 50 m) plot on the edge of the field.
The ancient Maya polity of Uxbenká is located on Santa Cruz lands, and the
village itself is nearly superimposed on the ancient city. The urban core of Uxbenká
covers an approximate area of 526 ha and comprises six plaza groups on leveled
ridgelines in the hilly landscape (see Figure 2.1). Group A contains the remnants of 23
carved sandstone stela dating to the Early and Late Classic periods and is presumed to be
the main ceremonial locus at the site (Prufer et al. 2011). Groups B-F are a contiguous
arrangement of plazas running along a ridgeline roughly 400m to the northwest of Group
A. The Group B plaza is a flattened hilltop and is surrounded by a series of range
structures, a large platform mound at its northern extent, and a ballcourt stands opposite
this at the south end of the plaza. A second ballcourt is located adjacent to the Group D
plaza.
Construction in Uxbenká’s urban core began in the Late Preclassic, with the
earliest known structure in Group A dating to 60 cal BC - cal AD 220 (Culleton et al.
2012). The massive effort of leveling and expanding ridgelines to form the Group B and
D plazas occurred slightly later, but still at the end of the Late Preclassic between cal AD
97
60-310 and cal AD 20-240, respectively. There was a flurry of replastering and plaza
renovation activity until the first part of the Early Classic Period, and then less evidence
for building activity between cal AD 350 and 550. Architectural modifications are
documented at Groups A, B, and D after AD 550, including extensive plastering of plaza
floors, laying paving stones, and the augmentation of facades on existing structures. The
latest dedicatory date preserved on stelae at Group A indicates that monument carving
continued until AD 781. Political disintegration and the abandonment of this city in the
Terminal Classic are topics of ongoing research at Uxbenká, but there is currently no
evidence for a Post-Classic (after AD 1000) occupation of the site.
Geoarchaeological investigations at Uxbenká provide evidence for early land-
clearing and erosion during the Middle Preclassic Period (ca. 970-620 cal BC), general
landform stability through the Classic Period (AD 300-800), and another episode of
erosion during the Terminal Classic (AD 800-900). This has been interpreted as the shift
from agricultural to urban land uses and back across the last several millennia (Culleton
et al., nd). There is little archaeological evidence for agricultural intensification in the
form of terraces or raised fields during the site’s history. This may be explained by the
capacity of the mudstone and sandstone bedrock to rapidly break down and form new
soils when exposed to weathering, a process that contributes to the “paradoxical” fertility
of the soils around Uxbenká and may play a role in the persistence of traditional swidden
cultivation in the region (Culleton et al. nd; Hartshorn et al. 1984:76-77). The physical
proximity of the Santa Cruz village to Uxbenká, the co-location of ancient and
contemporary land-uses, and the relatively low levels of intensification or technological
elaboration in the past and present farming systems provide a unique opportunity to
98
empirically estimate present day productive capacity of the land as a means to infer the
potential population supported at Uxbenká during its height in the Late Classic Period.
Methods
Arrays of 10x10 m sampling plots in planted milpas were selected in cooperation
with farmers and village representatives in a variety of settings around Santa Cruz in June
2009 and 2010. Slope and aspect of each plot were determined in the field with combined
compass and inclinometer. Slope was recorded in 5° increments and was converted to an
integer scale for regression analyses (e.g., 0-5° = 1, 5-10° = 2, etc.). UTM coordinates
were recorded with handheld GPS for integration with a GIS database. Working from a
digitized and orthorectified soils GIS basemap, each plot was also assigned a productivity
ranking based on Wilk’s (1981, 1991) classification of southern Belize soil types as
mapped and described by Wright et al. (1959). Following Wilk (1981), plots were ranked
on a scale from 0 (unusable) to 3 (good). Soil samples were collected from each plot at a
depth of 10-15 cm below the surface and analyzed for organic and inorganic carbon
content through loss-on-ignition (Dean 1974; Heiri et al. 2004) at the University of
Oregon, and for soil chemistry data (i.e., N, P, K, pH) at Oregon State University’s
Central Analytical Lab.
At the end of the growing season in late September and October, all corn within
each plot was broken by hand, and the number of suk’ub (Mopan: plantings) and
individual ears was counted. Bulk maize was weighed on a hanging scale to produce an
estimate of yield in kg/ha. During the first season, in a sample of roughly half the plots
(n=19 of 40), the corn was completely skinned and shelled, and the composition of edible
99
corn, waste corn, husk and cob were determined. Waste corn mainly results from weevil
infestation, rot due to fungus or bacteria, and sprouted corn. Usually only a small portion
of each ear was considered inedible, and was separated during shelling and fed to pigs.
The compositional data indicate a strong significant positive correlation between whole
ear weight (x) and yield of edible corn (y): Pearson’s r = 0.982; r2 = 0.965; p <0.00001; y
= 0.635x - 1.222. A less strong but also significant relationship is found between whole
ear weight and waste maize: Pearson’s r = 0.726; r2 = 0.527; p <0.0005; y = 0.099x +
0.138. Additional data from 2010, when all plot samples (n=40) were completely skinned
and shelled, bear out this relationship: edible maize, Pearson’s r = 0.955; r2 = 0.911; p
<0.00001; y = 0.601x - 0.077; waste, Pearson’s r = 0.415; r2 = 0.172; p = 0.008; y =
0.055x + 1.051. It is on this basis that bulk yields (kg/ha) are later converted into edible
yields (kg/ha) (Figure 4.1).
Examination of the bulk yields data indicates they were strongly dependent upon
planting density, or the number of suk’ub per plot (Pearson’s r = 0.722; r2 = 0.521; p
<0.0001; y = 41.45x + 540.74; see Figure 4.2). Planting density varies in the modern
setting for many reasons: shorter maize varieties can be planted closer together than taller
ones; maize intercropped with other plants (e.g., pepitorio) is more widely spaced to
reduce overshadowing; steeper slopes may be planted more densely if they are well
exposed; avoiding physical obstacles like unburned timber, rocks, or shallow soil may
force the plantings further apart; and, of course, using traditional sowing techniques (i.e.,
digging sticks and hand-casting seed) means that each farmer’s spacing differs based on
the length of his gait, willingness to negotiate physical obstacles, and other idiosyncratic
factors. The average number of suk’ub in the 100 m2 plots is roughly 55, or a planting
100
Good Corn
y = 0.6351x - 1.2224
R2 = 0.965
Waste Corn
y = 0.0994x + 0.1381
R2 = 0.5266
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0
Whole ear wt (kg)
She
lled
corn
wt (
kg)
Figure 4.1. Scatterplots of edible (circles) and waste (triangle) maize vs. whole ear weight for test plots on Santa Cruz milpas.
density of 5500/ha, while the majority range between 40 and 70 suk’ub per plot (4000 –
7000/ha).
To remove the effects of planting density, data were normalized by conversion to
yield per number of plantings (Tables 4.1 & 4.2). This indexed yield more closely reflects
the underlying productivity of the soil, and so regressions of other environmental
variables against this value are more likely to identify key causal variables. Although the
resulting unit kg/ha/suk’ub makes sense as a general productivity measure it is not easily
compared to other ethnographic data where yields are reported as production per unit area
(bushel/acre or kg/ha; e.g., as summarized by Barlow 2002:71). To compare yields
directly, values are normalized assuming an average density of 5500 suk’ub/ha.
101
y = 41.448x + 540.74
R2 = 0.5213
0
1000
2000
3000
4000
5000
6000
20 30 40 50 60 70 80 90 100 110 120
# suk'ub (plantings)
nul w
t (kg
/ha)
Figure 4.2. Relationship between bulk maize yield (kg/ha) and number of plantings per sample plot.
Table 4.1. 2009 bulk, edible and normalized maize yields
Lot# Bulk Maize
(kg/ha)
Edible corn
(kg/ha)
Normalized Yield (kg/ha/
suk’ub)
Lot# Bulk Maize
(kg/ha)
Edible corn
(kg/ha)
Normalized Yield
(kg/ha/ suk’ub)
19141 5630 3574 31.63 19153 1850 1174 16.30 19142 2390 1517 27.08 19154 2889 1834 33.96 19143 3400 2158 42.32 19155 2220 1409 21.34 19144 2830 1796 39.05 19156 1420 901 18.38 19145 1800 1142 30.05 19157 2720 1726 28.30 19147 1340 850 29.30 19158 2780 1764 36.01 19148 2500 1587 49.58 19159 2600 1650 40.24 19149 1960 1244 25.38 19162 2910 1847 37.69 19150 4050 2571 42.85 19163 1940 1231 34.19 19151 2400 1523 25.38 19164 2400 1523 33.11 19152 3350 2126 35.44 19165 2910 1847 34.20 19160 2110 1339 58.21 19167 300 189 7.01 19161 2450 1555 33.08 19183 3990 2533 58.90 19178 4280 2717 37.74 19184 4460 2831 30.78 19179 4960 3149 39.36 19185 3120 1980 37.36 19180 3360 2133 27.34 19186 3210 2037 42.45 19181 2594 1646 29.93 19187 2770 1758 41.86 19191 2390 1517 30.33 19188 2750 1745 34.91 19193 2450 1555 29.90 19189 3150 1999 32.25 19194 2030 1288 25.26 19190 2660 1688 24.83
102
Table 4.2. 2010 bulk, edible and normalized maize yields
Lot# Bulk Maize
(kg/ha)
Edible corn
(kg/ha)
Normalized Yield (kg/ha/
suk’ub)
Lot# Bulk Maize
(kg/ha)
Edible corn
(kg/ha)
Normalized Yield
(kg/ha/ suk’ub)
40701 4540 2740 38.06 40728 3640 1600 25.00 40702 3650 2100 34.43 40731 2580 1470 31.96 40703 1160 730 11.59 40732 3840 2040 37.78 40704 4700 3000 55.56 40733 2530 1630 31.96 40705 2650 1750 40.70 40734 4160 2560 49.23 40706 2580 1620 36.00 40737 3320 1900 29.69 40707 3900 2560 43.39 40738 3390 1920 45.71 40708 4090 2670 49.44 40739 3840 2540 44.56 40709 2630 1760 31.43 40740 2340 1420 25.36 40710 3390 2100 32.31 40741 3290 1780 34.90 40712 1330 800 14.55 40742 4580 2750 41.67 40713 2300 1530 24.29 40743 3590 1920 24.94 40714 2610 1730 31.45 40744 3640 2000 31.25 40721 3090 2000 32.26 40745 2840 1490 21.59 40722 3420 2110 30.58 40746 2730 1350 19.85 40723 1250 780 15.00 40748 1960 1180 20.00 40724 2030 1210 22.41 40749 2800 1580 27.72 40725 2490 1310 22.98 40750 4190 2700 42.19 40726 1720 890 14.83 40751 2420 1540 29.62 40727 2590 1640 26.89 40752 1880 1220 23.92
Comparing normalized whole ear weight for 2009 and 2010 maize yields show
similar mean and range values despite differing planting conditions (Figure 4.3). Farmers
in Santa Cruz considered 2009 a “bad” year for maize. The late onset of the dry season
combined with sporadic and heavy midday rains through May kept chopped vegetation in
milpas moist and difficult to burn. This was followed by dry conditions during the
summer (rainy) growing season. By October most farmers were breaking corn in earnest,
and many were clearing fields for an earlier start on the matahambre crop to make up for
anticipated shortfalls. By contrast 2010 had a more predictable termination to the dry
season and rains persisted through the summer rainy season. Farmers were less concerned
with breaking milpa crop or clearing matahambre early compared with the previous year
and more time and effort was instead devoted harvesting rice as a cash crop. Despite
these differences the mean yields between 2009 and 2010 are statistically
indistinguishable using a t-test with unequal variances (52.5 kg/ha/suk’ub in 2009 vs.
51.9 kg/ha/suk’ub in 2010; p = 0.870; Ruxton 2006). This sample is small and on-going
103
0
2
4
6
8
10
12
14
0 10 20 30 40 50 60 70 80 90 100
nul wt (kg/ha)/suk'ub
2009
2010
Figure 4.3. Comparison of 2009 and 2010 density-normalized maize yields. Means are statistically indistinguishable using a t-test with unequal variances (2009 = 52.5 kg/ha/suk’ub; 2010 = 51.9 kg/ha/suk’ub; p = 0.870).
work is focused on determining how the timing of the dry and rainy season rains may
impact crop yields. Rain and temperature gauges have recently been installed in Santa
Cruz village and are starting to provide quantitative meteorological data that can be
compared with land-clearing and planting schedules. The data currently available suggest
that a wide variety of variables act in concert to produce the observed more-or-less
normally distributed range of yields across the landscape and over time.
Spatial variability in the productivity of maize on lands surrounding Santa Cruz
and Uxbenká are detailed in the next section. I also compared edible maize yields in
kg/ha and density normalized yields in kg/ha/suk’ub against multiple environmental
variables (i.e., soil nutrients, pH, slope, aspect, and distance from Santa Cruz village
Figures 4.4 and 4.5). Regressions made against log-transformed, density-normalized data
show no correlations between yields and these variables for each year with the exception
of a weak correlation between K (potassium) and yield in 2010 (Tables 4.3 and 4.4).
104
Figure 4.4. Scatterplots of bulk maize yields (black circle: 2009; grey circle: 2010) vs. environmental variables, showing the range of scatter and lack of correlation.
Figure 4.5. Scatterplots of density-normalized maize yields (black circle: 2009; grey circle: 2010) vs. environmental variables, showing the range of scatter and lack of correlation.
105
The two-year aggregated data show also show a weak correlation between K and yield
(Table 4.5).
Table 4.3. Regression of 2009 log-transformed maize yield normalized for planting density with respect to environmental variables.
Edible maize (kg/ha/suk’ub) vs. Pearson's r r2 p-value regression equation
Distance from Santa Cruz (km) 0.227 0.051 0.159 y =0.29x + 1.60
Slope Index 0.172 0.030 0.287 y = -0.10x + 1.73
%OC 0.073 0.005 0.656 y = 0.11x + 1.79
%CO3 0.042 0.002 0.803 y = -0.12x + 1.52
pH 0.015 <0.001 0.928 y = 0.08x + 1.64
K (ppm) 0.005 0.003 0.753 y = 0.05x + 1.56
TKN (ppm) 0.017 <0.001 0.916 y = 0.02x + 1.63
TP (ppm) 0.048 0.002 0.764 y = 0.06x + 1.53
Table 4.4. Regression of 2010 log-transformed maize yield normalized for planting density with respect to environmental variables.
Edible maize (kg/ha/suk’ub) vs Pearson's r r2 p-value regression equation
Distance from Santa Cruz (km) 0.122 0.015 0.451 y = 0.13x + 1.64
Slope Index 0.207 0.043 0.201 y = 0.13x + 1.62
%OC 0.073 0.005 0.658 y = 0.12x + 1.78
%CO3 0.264 0.070 0.104 y = 0.44x + 2.35
pH 0.216 0.047 0.179 y = 1.47x + 0.51
K (ppm) 0.403 0.162 0.010 y = 0.54x + 0.35
TKN (ppm) 0.096 0.009 0.556 y = 0.15x + 1.14
TP (ppm) 0.287 0.082 0.073 y = 0.36x + 0.76
Table 4.5. Regression of 2-year aggregated log-transformed maize yield data normalized for planting density with respect to environmental variables.
Edible maize (kg/ha/suk’ub) vs Pearson's r r2 p-value regression equation
Distance from Santa Cruz (km) 0.173 0.030 0.125 y = 0.19x + 1.63
Slope Index 0.031 <0.001 0.783 y = 0.02x + 1.68
%OC 0.082 0.007 0.473 y = 0.13x + 1.80
%CO3 0.187 0.035 0.103 y = 0.36x + 2.23
pH 0.094 0.008 0.403 y = 0.56x + 1.25
K (ppm) 0.231 0.053 0.039 y = 0.27x + 1.01
TKN (ppm) 0.048 0.002 0.669 y = 0.07x +1.45
TP (ppm) 0.187 0.035 0.097 y = 0.24x + 1.07
Maize Yields and Modern Populations
Geographic Information Systems (GIS) were used to interpolate the productivity of the
lands surrounding Santa Cruz and Uxbenká (bulk yields normalized to planting density).
106
The locations of milpas sampled in 2009 and 2010 were plotted in ArcGIS 10.0 along
with archaeological settlements and the architecture within the Uxbenká site core. The
normalized yield or yield index (kg/ha/suk’ub) was interpolated between plots with the
Spatial Analyst toolset using a Nearest Neighbor method that assigns values to locations
based on the surrounding measured values. Each 18.5 m 18.5 x 18.5 m (0.034225 ha) cell
has an associated yield index value ranging from 11.8 – 92.9 kg/ha/suk’ub, with an
average value of 47.9 kg/ha/suk’ub (Figure 4.6). The distinction between the more fertile
box lu’um and the less favorable chik lu’um soils influences the raster in two ways. The
general north to south gradient of greater to lesser maize yields does map on to the known
distribution of the soils in Santa Cruz described earlier, with the chik lu’um located
primarily between the village and the Rio Blanco. At the same time, there are relatively
few maize plots on the chik lu’um because milpas in this zone are typically planted in dry
rice and samples are therefore difficult to obtain. A t-test assuming unequal variances run
on the 3 chik lu’um samples vs. the majority of the plots positioned to the north of the
village on box lu’um soils (n = 75-77 depending on the variable) does show a significant
difference between the sample means in % organic carbon (p < 0.002; chik lu’um mean =
9.2%; others = 14.2%). Further soil sampling will be directed towards the area south of
the village in future field seasons.
The exact political boundaries of Santa Cruz are in the process of being
determined by local community leaders, but Santa Cruz village lands are estimated to be
approximately 16.08 km2 for the purposes of this study based on the recent history of
land use practices. Milpas are not planted in a buffer zone of ~0.5 km around the village,
corresponding to the range that domestic pigs will travel to forage. Removing this 1.63
107
Figure 4.6. Interpolated raster of maize yields around Santa Cruz village (image created by C. Ebert).
km2 area from analysis, total arable land available to farmers in Santa Cruz is estimated
to be 14.45 km2. The total area of Santa Cruz land cultivated in 2009 and 2010 was
estimated from a satellite image (Worldview II) taken in April 2010 that covers a 100
km2 around Santa Cruz/Uxbenká. This image provides ~60 cm resolution and is
composed of 8 multispectral bands that include Red Edge (705 - 745 µm) and Near
infrared (IR) bands (760-900 µm). The color IR image (including the red edge band) was
used for photo interpretation due to its broader spectral resolution that allowed cleared
agricultural plots to be distinguished from the surrounding vegetation. A total of 134
fields were identified and digitized for both years totaling 209.88 ha, an average of
104.94 ha of land cultivated each year. The total cultivated land comprises both milpa
and matahambre cultivation of subsistence crops such as maize, beans, and ground foods,
108
as well as cash crops including dry-field rice and seed pumpkins (pepitorio). A total of
111 milpas fall into the raster area and represent a total area of 168.52 ha.
As discussed earlier, the yield index is a normalized value that accounts for the
influence of planting density on potential yields expressed as kg/ha/suk’ub. The average
yield index for the raster is 47.86 kg/ha/suk’ub. Taking representative values of 4000,
5500, and 7000 suk’ub/ha (see above), the yield index can be converted to bulk yield
(kg/ha) to model low, medium, and high yield scenarios. Note that for heuristic purposes
these scenarios could also be used to approximate variation caused by weather
conditions, pest activity, theft, etc.
To get a sense of how realistic the interpolation and average yield index might be
for estimating the ancient population at Uxbenká, I converted the yield index into
absolute yields based on the area currently cultivated by farmers (i.e., 104.49 ha) to
compare these modeled results with the census data from Santa Cruz village. Absolute
yields in kg were converted to edible corn using an empirically-derived conversion from
whole ear weight to kernel weight (0.60), and then multiplied by 0.95 to account for the
estimated difference in equilibrium moisture content (EMC) for October maize in Belize
(~19% w/w) versus the dry weight EMC of stored US maize (~14% w/w). Absolute
yields of dry corn are converted to yield in kcal assuming 3650 kcal/kg for dry maize
(USDA Nutrition Database).
Positing an average daily caloric requirement of 2500 kcal/day for Santa Cruz
villagers based on adult caloric needs is a simplifying but conservative assumption when
calculating population. Working from FAO/WHO/UNU (1991) estimates, 2500 kcal/per
day is the rough average of an adult male subsistence farmer (2780 kcal/day; FAO et al.
109
[1991], Table 10) and a ‘rural woman in a developing country’ (2235 kcal/day; FAO et
al. [1991], Table 14). Assuming an equal sex ratio this average person leads to a
relatively conservative population figure, as children and the elderly have lower caloric
demands in general.
Based on these model assumptions I estimate the village population at between
~460 and 800 people, higher than the current census of 400 people (Wainwright 2007;
Table 4.6). Since not all cultivated land is devoted to subsistence, and not all subsistence
crops yield caloric returns equal to maize, these figures can be corrected assuming
different proportions of land devoted to maize (Table 4.7). Additional work is needed to
determine the percentage of maize consumed on average, but an estimate of 70-80%
maize cultivation seems reasonable based on informal observation in Santa Cruz
throughout the year and considering that two crops of maize are grown per year and dry
field rice, beans and ground foods tend to be grown less frequently. Assuming 70-80%
maize cultivation, then ~105 ha would support a village population closer to ~440-
500people (per year). Taking the total arable land around Santa Cruz as 1445 ha, this
suggests an average of 7.2% of land cultivated each year and an average fallow period of
13.8 years. This average for the entire area appears reasonable considering some fields
are routinely cleared every 5 years, and more distant forest stands can remain in fallow
for 20 years or more (e.g., some higher stands of forest at the foot of the rock patch, and
section to the northwest towards Ya’ax Ha).
110
Table 4.6. Estimated annual maize yield and potential population of Santa Cruz village (values rounded for clarity)
Planting Density
(suk'ub /ha) Gross Yield (kg) Edible Corn (kg) Dry weight
(kg) Total Energy
(kcal) Population @ 2500 kcal/day
4000 200910 120540 114520 417985130 460 5500 276250 165750 157460 574729550 630 7000 351590 210950 200400 731473980 800
Table 4.7. Estimated potential population of Santa Cruz village assuming % of land devoted to maize (values rounded for clarity)
Population @ 2500 kcal/day Planting Density
(suk'ub /ha) 60% Maize Cultivation
70% Maize Cultivation
80% Maize Cultivation
90% Maize Cultivation
100% Maize Cultivation
4000 270 320 370 410 460 5500 380 440 500 570 630 7000 480 560 640 720 800
Population Estimates for the Uxbenká Polity
Without the current political boundaries limiting the available arable land and no
evidence of ancient political boundaries, the potential catchment for Uxbenká can be
modeled as a series of concentric rings radiating out from the site core. In this study the
catchment area is centered on Group B (Structure B1; Figure 4.7) and arable land area
was calculated in 1 km radii subtracting the 526 ha site core that was probably not
cultivated with maize crops during the Classic Period (Chapter III; Culleton et al. nd).
These concentric rings were also truncated at the edge of the high karst “rock patch” to
the south because this rugged terrain is not suitable for agriculture. The interpolated yield
raster does not cover the entire extent of land potentially under cultivation by the
inhabitants of Uxbenká, so the average index value is assumed for the entire area.
Estimated potential population for each catchment is presented in Table 4.8, assuming
low, average, and high planting density and varying the proportion of maize cultivation.
These values are calculated as person-years, or the number of people that could be fed for
a year if the entire area was put under cultivation.
111
Figure 4.7. Hypothetical catchments centered on Uxbenká used to estimate the total maize production for each additional 1 km radius (image created by C. Ebert).
Since these values are time-dependent, the maximum carrying capacity at
Uxbenká can be calculated under varying fallow cycles. For this calculation, assume that
75% of the land is devoted to maize cultivation, and for the sake of simplicity assume
similar yields at each fallow length. The population at each fallow level is presented in
Table 4.9, and ideally represents the population that could be sustained indefinitely at a
given intensity of land clearing. These numbers are also translated into number of
households, assuming an average of 5 persons per household, a commonly applied
estimate for Mesoamerican nuclear families (e.g., Culbert and Rice 1990, and papers
therein; Webster et al. 2000).
112
Table 4.8. Estimated maximum population for Uxbenká assuming different catchment areas (values rounded for clarity).
3 km (2774.79 ha) person-yr @ 2500 kcal/day
Suk'ub/ha kcal % maize 50 60 70 80 90 100
4000 10910590890 5980 7170 8370 9570 10760 11960
5500 15002062470 8220 9860 11510 13150 14800 16440
7000 19093534050 10460 12550 14650 16740 18830 20920
4 km (4613.34 ha) person-yr @ 2500 kcal/day
Suk'ub/ha kcal % maize 50 60 70 80 90 100
4000 17829013530 9770 11720 13680 15630 17580 19540
5500 24514893600 13430 16120 18810 21490 24180 26870
7000 31200773670 17100 20520 23930 27350 30770 34190
5 km (6633.03 ha) person-yr @ 2500 kcal/day
Suk'ub/ha kcal % maize 50 60 70 80 90 100
4000 25847963050 14160 17000 19830 22660 25500 28330
5500 35540949190 19470 23370 27260 31160 35050 38950
7000 45233935340 24790 29740 34700 39660 44610 49570
6 km (8948.46 ha) person-yr @ 2500 kcal/day
Suk'ub/ha kcal % maize 50 60 70 80 90 100
4000 35005955910 19180 23020 26850 30690 34530 38360
5500 48133189370 26370 31650 36920 42200 47470 52750
7000 61260422840 33570 40280 46990 53710 60420 67130
Discussion
The results suggest that within the area of arable land potentially under the
political influence of Uxbenká, i.e., the area within 6 km (to a point equidistant from
Lubaantun, the nearest regional center during the Classic Period) and excluding the karst
ridge and lands to the south, ~7500-13,000 people could have been supported on a 5-year
fallow cycle (Table 4.9). At longer fallow cycles requiring more available land the
potential population is proportionately less. Wilk (1984, 1991) assumed that land cleared
for swidden cultivation would need 30 years of fallow to return to high (primary) forest
113
Table 4.9. Estimated population of Uxbenká at various catchments and fallow length assuming 75% maize cultivation (values rounded for clarity).
3 km (2774.79 ha) Population at Fallow (yr) Households (5 person/family)
Suk'ub/ha 5 10 15 20 30 5 10 15 20 30
4000 2390 1200 800 600 400 480 240 160 120 80
5500 3290 1640 1100 820 550 660 330 220 160 110
7000 4180 2090 1400 1050 700 840 420 280 210 140
4 km (4613.34 ha) Population at Fallow (yr) Households (5 person/family)
Suk'ub/ha 5 10 15 20 30 5 10 15 20 30
4000 3910 1950 1300 980 650 780 390 260 200 130
5500 5370 2690 1790 1340 900 1070 540 360 270 180
7000 6840 3420 2280 1710 1140 1370 680 460 340 230
5 km (6633.03 ha) Population at Fallow (yr) Households (5 person/family)
Suk'ub/ha 5 10 15 20 30 5 10 15 20 30
4000 5670 2830 1890 1420 940 1130 570 380 280 190
5500 7790 3890 2600 1950 1300 1560 780 520 390 260
7000 9910 4960 3300 2480 1650 1980 990 660 500 330
6 km (8948.46 ha) Population at Fallow (yr) Households (5 person/family)
Suk'ub/ha 5 10 15 20 30 5 10 15 20 30
4000 7670 3840 2560 1920 1280 1530 770 510 380 260
5500 10550 5270 3520 2640 1760 2100 1060 700 530 350
7000 13430 6710 4480 3360 2240 2690 1340 900 670 450
based on a model of sustainable village size for K’ek’chi farmers in southern Belize.
Fallow times short of that were assumed to lead inevitably to declining yields over the
longer term as soil nutrients become depleted and the spread of grasses and weeds inhibit
the re-establishment of arboreal species. Wilk found that this eventually forced people to
use more intensive cultivation strategies via increased labor or soil augmentation (e.g.,
fertilizer). However, if uncleared forest was available the hypothetical village could
relocate and start anew elsewhere.
Although some aspects of Wilk’s heuristic model are not directly analogous to the
Uxbenká case (particularly the assumption that sub-climax conditions are inherently
114
unsustainable; cf. Hartshorn et al. [1984] and the “paradoxical fertility” of the Toledo
Beds) these assumptions provide a starting point to model the population size at the 30-
year fallow cycle. Assuming a 3 km radius catchment and average planting density, these
lands could support a hypothetical village of roughly 550 members. A 5 km catchment (a
default village catchment size in Wilk’s analysis) could support a population of 1300. In
the contemporary setting, village populations in the area range between 300 and 500
people (excluding the largest Maya town of San Antonio) and catchments are closer to a
3-4 km radius. Fallows are also much shorter than 30 years. Given that the sustainability
of farming systems should be considered over generations, it is difficult to say that the
shorter fallows observed today will lead to the negative consequences predicted by
Wilk’s analysis for subsistence farmers in the region. The population estimates made here
will inevitably be improved by long-term data collection designed to establish the linkage
between length of fallow and productivity. These data will be required to make the model
more dynamic and applicable to analyzing diachronic processes.
The population estimates presented here give a sense of what level of intensified
food production may have been practiced in the Uxbenká environs. Because these
population estimates are based on maize yields per area, they indirectly assume a constant
population density per area for any given planting density and fallow length. For a 5-year
fallow as shown in Table 4.9, planting density of 4000 suk’ub/ha supports a population
density of 65.5/km2, 5500 suk’ub/ha supports 90.0/km2, and 7000 suk’ub/ha supports
114.6/km2, regardless of size of the catchment area under consideration (though longer
fallows would decrease the population density by decreasing overall production at a
given planting density).
115
These figures, which do not include calories and nutrients from other crops or
wild foods may well be conservative. Nonetheless, they are generally consistent in
magnitude with, if somewhat lower than, several of the broader areal population densities
based on residential structures from Rice and Culbert’s (1990) summary for Lowland
Maya centers (see Table 4.10): e.g., the Copan Valley in total, 43.2/km2 (Webster and
Freter 1990); rural areas within 10km of Tikal, 153.1/km2 (Culbert et al. 1990);
Guatemalan lake basins and the Yaxha Polygon, 163.2 – 260/km2 (A. Chase 1990);
Nohmul, 150.5/km2 (Pyburn et al. 1990).
Translating the Uxbenká population densities to household densities assuming 5-
persons per household gives a range of 13.1 – 22.9 households/km2, with each household
having an average of 4.4 - 7.6 ha for cultivation over the long term (assuming some form
of usufruct land tenure). The current settlement survey indicates that most of these
households would have been located on hilltops or extended across ridgelines throughout
the hilly and steeply incised landscape. Assuming a 5-year fallow and an annual plot size
of 1.5 ha (as with the contemporary situation) ancient households at this settlement
density would be on the cusp of choosing between planting in more distant outfields to
acquire more land, decreasing fallow time (e.g., a 1.5 ha/yr in a 3 yr fallow over 4.5 ha),
or increasing the planting density on the same amount of land. These are among the
simplest and presumably earliest strategies employed along the spectrum of agricultural
intensification (in the sense of Boserup 1965). They are strategies that would leave little
trace in the archaeological record. Low-level agricultural intensification is in general
accord with the lack of evidence for substantial soil management features at Uxbenká,
those that would signal large labor inputs to mitigate declining productivity or would put
116
Table 4.10. Population estimates for Lowland Maya sites, after Rice and Culbert (1990: Table 1.3)
Site Area (km 2)
Estimated Population
Density (range, pop/km 2)
Density (pop/km 2)a
Late Preclassic Period Seibal
Center 1.6 1644 1027.5 Peripheries 13.6 7974 586.3
Total 15.2 9618 632.8 Komchen 2 2500-3000 1250.0-1500.0 1375.0 Late Classic Period Copan
Urban core 0.6 5797-9464 9661.7-15773.3 12717.5 Copan pocket, rural 23.4 9360-11,639 400.0-497.4 448.7
Outside Copan pocket, rural 476 3010-3725 6.3-7.8 7.1
Copan Valley, Total 500 18,417-24,828 36.8-49.7 43.2 Quirigua (center) 3 1183-1579 394.3-526.3 460.3 Tikal
Central 9km2 9 8300 922.2 Next 7km2 7 4975 710.7
Remainder within boundaries 104 45720 439.6
Total within boundaries 120 62000 516.7 Rural within 10km 194 29696 153.1
Macanche-Salpeten Basin 27.9 7262 260.3 Yaxha-Sacnab Basin 29.5 6253 212.0 Quexil-Petenxil Basin 23.5 3836 163.2 All lake basins 78.3 17351 221.6 Yaxha Polygon 237 42047 177.4 Tayasal
Spine 8 6861-10,400 857.6-1300.0 1078.8 Outer Ring 18 7719-11,000 428.8-611.1 520.0 Periphery 64 7371-11,172 115.2-174.6 144.9
Total 90 21,951-32,272 243.9-358.6 301.2 Late/Terminal Classic Nohmul 22 3310 150.5 Sayil (by mounds) 3.4 8148-9990 2396.5-2938.2 2667.4 Sayil (by chultuns) 3.4 4900-10,000 1441.2-2941.2 2191.2 Late Postclassic Santa Rita 5 4958-8722 991.6-1744.4 1368.0
a: Mean density is given for sites with a range of estimates.
more proximate marginal land into production (e.g., terraces, raised fields, or simply
demarcated fields; Culleton et al. nd). More work on both the food production system and
the settlement archaeology remains to be done but these initial population estimates
117
suggest that the area conceivably under the political influence of Uxbenká could have
supported 7500-13,000 people without resorting to archaeologically obvious intensive
agriculture strategies.
Classic Period Lowland Maya city centers are thought to support densities
between 6 and 100 times the average for the broader landscape (Culbert and Rice 1990),
the latter extreme representing architectural intensification that most contemporary city
dwellers would have no trouble recognizing as urban. Investigations characterizing the
residential nature of the Uxbenká site core are underway, but the current data from survey
and excavations suggest that it is much closer to the lower end of the urban density
spectrum. Assuming 5 times the average population density with a 5 year fallow in an
area of 0.526 km2, gives an estimate of ~237 people living in the site core, or ~475 at 10
times the density. Based on Webster’s (1985; Webster et al. 2000) assumption of a
maximum of 10% of ancient Maya populations being elites and specialists (i.e., those not
involved in food production; at most 5% belonging to each group), and a population of
10,550 for the 6 km radius around Uxbenká at a 5-year fallow, we derive a non-
producing population of 1055 people at Uxbenká at its height. If only the elite segment
resided in the site core, this gives a maximum estimate of ~525 people, which is
reasonably close to the larger estimate of 475 based on relative population density.
Translating elite population estimates into numbers of households is less straight
forward than for the broader population of Uxbenká because of differences in the ways
elite households were constituted as social and economic entities. Polygyny among elite
families is well-attested, and the inclusion of retainers, specialists, and slaves could
increase the houshold size considerably (Webster et al 2000: 158-160, 165). Working
118
from the assumption of a 5-person household, the two density-based estimates for the site
core population translate to 47 to 95 households in the core, respectively, and the elite
proportion estimate suggests a maximum of 105 households. Using a hypothetical
average household size of 20, the core would have been composed of ~12 to 24
households, or 26 households using the elite proportion of the overall population. These
estimates can be developed into testable predictions about the number and types of
structures that should be found by ongoing household investigations in the site core,
keeping in mind that “elite” structures and burials cover a spectrum from modest to
elaborate (Webster et al. 2000: 165). Results of those studies will provide an independent
test of the assumptions involved in this population reconstruction, and highlight specific
areas for revision and refinement.
Conclusions
The agricultural productivity of the present-day landscape was used to estimated
the maximum potential population size for the ancient Maya center of Uxbenká. Maize
yields in milpas planted by farmers around the village of Santa Cruz were quantified
during the 2009 and 2010 harvest seasons, and compared with environmental variables
including soil nutrients (e.g., N, P, K, pH, organic and inorganic carbon) and landscape
attributes (e.g., slope, aspect, distance from the village). Maize yields were found not to
correlate with measured variables, except for a very weak positive correlation with
distance from the site core. Planting density, which varies with the type of maize planted,
was found to heavily influence yields and is dependent upon intercropping with other
cultivars and the presence of physical obstacles in cleared milpas. The lack of correlation
119
between yields and a range of environmental variables is consistent with other
ethnographic studies on maize production that suggest a range of confounding factors
(e.g., soil, weather, maize variety, pests, and farming experience) ultimately dictate the
outcome at harvest.
Yield values were controlled for planting density and incorporated into a
geospatial database to interpolate a productivity raster of the lands surrounding Uxbenká.
Taking the average maize yield per area and assuming daily caloric needs for ancient
inhabitants, the maximum sustainable population of the Uxbenká polity during the
Classic Period is estimated to be between 7500 and 13,000 people within a 6km radius.
This population is modeled at a five-year fallow period, just on the cusp of a short fallow
system suggestive of a low level of agricultural intensification. The lack of
archaeological evidence for intensive farming strategies (e.g., terracing, field
demarcation, irrigation systems) in the vicinity of Uxbenká is consistent with this model
result. Assuming the elite population resided in the urban core of the site and that it was
5% of the total population, the model predicts the presence of ~525 elites, though the
number of elite households is difficult to reliably estimate because of their unique social
and economic makeup. Productivity-derived predictions of population size and household
density within the ancient Uxbenká polity provide expectations for the material record
that can be tested through future work in household, settlement and landscape
archaeology.
120
CHAPTER V
THE IDEAL FREE AND DESPOTIC DISTRIBUTIONS AND ANCIENT MAYA
SETTLEMENT AT UXBENKÁ, BELIZE
This chapter was prepared as an unpublished co-authored manuscript with Dr.
Winterhalder, Ms. Ebert, Mr. Ethan Kalosky, Dr. Prufer, and Dr. Kennett. I conceived of
the settlement models that incorporate data on productivity derived from modern maize
yields (see Chapter IV), hydrology and proximity to the site core, and also conducted
settlement excavations and ceramic analyses at two settlement groups to expand the
chronological dataset. I processed the radiocarbon dates that form the overall settlement
chronology. Dr. Winterhalder provided guidance on the application of the Ideal Free and
Despotic Distributions and gave valuable feedback on the implementation and
interpretation of these models. Ms. Ebert summarized data on productivity, hydrology
and proximity for each settlement group catchment using the GIS database, as well as
providing overall map coverages and elements of key figures. Mr. Kalosky directed and
conducted much of settlement survey and mapping that forms the settlement database for
the analysis presented here, and graciously shared preliminary results of a least-cost path
analysis of the project area. Dr. Prufer oversaw the original settlement field work and
provided access to the settlement and chronological data, and provided useful discussion
on the interpretation of the model results. Dr. Kennett helped with organization and
presentation of the data and the model design, as well as providing critical feedback on
interpretive aspects of the models.
121
The florescence of ancient Maya culture from the Late Preclassic through the
Classic Period was marked by increased social differentiation and institutionalized status
hierarchy, agricultural intensification, elite control of water resources, expanded trade
and exchange, interpolity conflict, organized warfare, and environmental degradation
(Demarest 2004; Fedick 1996a; Lentz 2000; Scarborough 2003; Schele and Freidel 1990;
Webster 2002). The processes of polity formation, settlement expansion, and political
decline in the ancient Maya Lowlands involved the dynamic interaction of social and
ecological factors influencing each other on multiple spatial scales, from the broadest
scale of political cooperation and conflict between multiple polities, to smaller scales of
interaction between factions or even individual commoner households within polities.
The connection between population increase, intensive food-production, and
environmental degradation is central to ecologically based explanations of the emergence
and decline of ancient Lowland Maya societies. However, many explanatory narratives of
the rise and fall of Maya polities take for granted that one or more of these processes is
operating without demonstrating it, or take evidence of one as a proxy for the others.
Demarest’s (2004:258, Figure 10.10) causal model for the collapse of Late Classic
Petexbatún, for example, placed population growth during the Late Preclassic as a prime
mover that also influenced the shift to shorter and shorter fallow times during the Classic
Period. He argued that increased intensification led to environmental degradation and
undermined the resource base that was rapidly overshot by a growing population. This
resulted in increased warfare for prime agriculture lands, social upheaval, and settlement
disruption through immigration and abandonment.
122
Cowgill (1975a, 1975b) has cogently argued against the assumption of intrinsic
population growth and the inevitable response of agricultural intensification (contra
Boserup 1965), but it is clear in the Maya region that population densities were greater at
many centers during the Classic Period compared with 2000 years earlier in the
Preclassic. There are also plausible causal linkages between population density, land-use
practices, resource availability and social behaviors of household settlement and
production that can be empirically demonstrated. So attempting to develop coherent
models of the consequences of changing population densities within an ecological
framework and applying them to archaeological data is a reasonable theoretical endeavor.
A set of models developed in Human Behavioral Ecology provide a framework
that incorporates explicit relationships between population density, habitat quality and
human decision-making that can be used to investigate the dynamic process of settlement
expansion. Specifically, the Ideal Free Distribution and related Ideal Despotic
Distribution (Fretwell 1972; Fretwell and Lucas 1969; Sutherland 1996) show great
potential for exploring the causal connections between socioecological conditions and
human behavior at multiple spatial and temporal scales. The scalar flexibility of these
models makes them particularly well-suited to addressing archaeological problems on
local and regional scales over decades, centuries or millennia.
The Ideal Free and Despotic Distributions
The Ideal Free Distribution (IFD) is a formal habitat choice model developed in
population ecology that incorporates density-dependent and density-independent
environmental factors of habitat suitability to generate testable predictions about
123
settlement behavior (Fretwell 1972; Fretwell and Lucas 1969; Sutherland 1996). The IFD
assumes that all members of a population are equal competitors for resources, have equal
ability to evaluate all available habitats (which implies a sort of theoretical omniscience
of the landscape for individuals), will always choose the most suitable habitat to settle
(the ideal of the IFD), and are able to relocate to any habitat at will (the free of the IFD).
Settlement locations or habitats are ranked by their relative suitability, a summary of
overall resource richness within a given area (Figure 5.1A). The IFD predicts that the
most suitable habitat (H1 in Figure 5.1A) is occupied first, and as population grows,
suitability in this habitat drops due to density-dependent resource depletion or
interference arising from competition. When suitability declines to that of the second-
ranked resource patch H2 at population density A, further population growth will be
divided between them. This process continues as population density increases and habitat
suitability declines to that of the lowest ranked habitat H3 at population density B. The
tempo and mode of this process may be affected by changes in suitability that affect all
habitats (e.g., climate change, adoption of novel technology, etc.). Another variation of
the IFD includes the Allee effect (Figure 5.1B), in which habitat suitability initially
increases with population density. Typical examples of the Allee effect in human groups
include: greater availability of suitable mates; increased food production due to collective
effort in construction and maintenance of raised fields, terraces, or irrigation systems; and
better opportunities for collective defense of resources. In either version, an equilibrium
population distribution is achieved between all habitats.
The freedom of any individual to relocate to a more favorable habitat at will is
conceivable among groups at relatively low regional population densities and high
124
Figure 5.1. Habitat rankings under assumptions of A) the Ideal Free Distribution, and B) with the Allee effect (after Kennett et al. 2006, 2008; Sutherland 1996).
residential mobility, but less likely in more socially stratified or territorially
circumscribed populations. A variation of the IFD that assumes unequal competitive
advantage, and hence the ability of some individuals (or groups of individuals) to exclude
others from a habitat, is the Ideal Despotic Distribution (IDD). Under the IDD individuals
still seek to settle in the highest ranked (i.e., ideal) habitat, but the presence of groups
with competitive advantage prevent immigration to these habitats. This has the effect of
mitigating density-dependent declines in habitat suitability within the best habitats, and
pushes others into lower ranked areas sooner than predicted by the IFD. In contrast to the
IFD, when IDD conditions exist, the process of competitive exclusion leads to an
equilibrium population distribution with disproportionately greater population densities in
lower ranked habitats. Such distributions are familiar to contemporary urban dwellers –
the favelas outside of Rio de Janeiro can be considered an extreme example of this
outcome – and so commonplace that one’s intuitive sense might be that IDD conditions
are a more likely default expectation than IFD conditions in human settlement. In fact,
125
the IFD is often taken as a null hypothesis in population ecology models against which
the effects of competition and unequal access to resources can be measured (Kennett and
Winterhalder 2006:89; Sutherland 1996).
The IFD and IDD are sufficiently general in formulation to allow them to be
adapted to a broad range of social and environmental settings to make predictions about
the processes of settlement and resource exploitation in past and present human
populations. “Habitat suitability” is an index of all social and ecological variables that
could bear on individual fitness (however that is conceived) and therefore can be defined
for specific research questions informed by a knowledge of relevant ecological variables,
mode of food production, and degree of technological complexity or status
differentiation. This flexibility makes the IFD and IDD amenable to starting with very
basic models that incorporate one or two key variables (e.g., access to water and
abundance of shellfish beds among coastal hunter-gatherers in an arid environment),
testing the model predictions against archaeological observations, and then refining the
concept of habitat suitability to include other predictive social or ecological variables. In
this way, developing and testing an IFD model iteratively can serve as a tool for
identifying relevant variables that have not been recognized or fully accounted for. Or, as
noted above, the failure of an IFD model to predict the observed distribution of
settlements may indicate the presence of interference competition, suggesting an IDD
condition prevails, and thereby focusing research on explaining the emergence and
maintenance of despotic conditions.
Another aspect of the IFD and IDD models that makes them productive for
archaeological inquiry is their dynamic and diachronic formulation, which opens up
126
useful avenues for explaining both stability and change in the archaeological record over
long periods. Many models in the evolutionary sciences begin as thought experiments
premised upon a time-transgressive narrative where competition under certain conditions
leads to a specific set of Evolutionarily Stable Strategies (ESS) for individuals, such as
classic models like the Prisoner’s Dilemma, Hawks vs. Doves, and so on (Kennett 1998;
Kennett and Clifford 2004; Smith 2000; Smith and Winterhalder 1992). The implied time
scale under which these strategies evolved (i.e., in the literal sense of the biological
evolution of innate behavioral tendencies, not the figurative or metaphorical usage of
evolution as any change or development in a group or individual) is typically assumed to
be on the order of >105 years in the Environment of Evolutionary Adaptiveness (EEA)
and to have already resulted in what we observed today as the distinct set of ESS for a
given species in its habitat. There is rarely a sense that the initial stages of the process
could ever be observed directly among living populations, except as they are
recapitulated by undergraduate test subjects in evolutionary psychology laboratories
around the world.
Similarly, for population biologists the diachronic aspect of the IFD and IDD
largely serves to provide a framework for understanding the equilibrium (or non-
equilibrium) population distributions observed in the present, i.e., the synchronic view of
a target population studied in the field. The scenario described by the IFD and IDD is
essentially the colonization of an unoccupied habitat by a novel species, a process
difficult to observe and describe over short timescales of years or decades of fieldwork,
but one that is represented in the centennial- to millennial-scale archaeological records of
much of the world. Assuming an adequately sampled and temporally-resolved record,
127
archaeologists have the opportunity to consider the diachronic aspects of the IFD and
IDD, as well as to take synchronic snapshots at specific points in culture history to
explore abrupt and discontinuous processes of societal change such as the development of
new technologies and food production techniques, resource intensification, migration,
colonization, and the emergence of social inequality.
Archaeological Applications of the IFD and IDD
Most applications of IFD and IDD models to archaeological problems have been
carried out by D.J. Kennett, B. Winterhalder, and their colleagues, primarily applied to
hunter-gatherer groups on California’s Northern Channel Islands (Kennett 2005:32-36,
229-233; Kennett et al. 2009; Winterhalder et al. 2010), to agricultural societies in
Polynesia (Kennett and Winterhalder 2008; Kennett et al. 2006), and to diachronic
patterns of trade and interactions between coastal and island populations along the west
coast of North America (Fitzhugh and Kennett 2010). Although these case studies are
largely in island contexts at scales ranging from individual islands (e.g., Rapa) to small
nearshore groups (e.g., California’s Channel Islands) to multiple and geographically
dispersed groups (e.g., Polynesia), the IFD and the IDD are equally applicable to
mainland continental settings as an early application of the IFD model to colonization
and expansion of Neolithic populations in Spain has demonstrated (McClure et al. 2006).
On California’s Northern Channel Islands, the IFD and IDD models have been
articulated with principles of Central Place Foraging theory (which guides habitat
definition by characterizing the size and content of resource patches; Orians and Pearson
1979; Stephens and Krebs 1986) to explore the process of settlement expansion through
128
the Holocene in terms of resource intensification, technological innovation, and emergent
social inequality. In the earliest work, predictions of an IFD model for the Channel
Islands that evaluated settlement location (habitat suitability) in terms of access to fresh
water (using drainage size as a proxy), extent of shellfish-rich rocky intertidal zones
foraging locales, and area of kelp forest for fishing, indicated that the earliest occupation
sites should be located at the mouths of the largest drainages on the islands and these
should also host most persistent settlements (Kennett 2005). Early and persistent
settlements at Arlington Canyon, Cañada Verde, Lobo Canyon, and Old Ranch Canyon
on Santa Rosa, and Central Valley and Prisoner’s Harbor on Santa Cruz conform to these
predictions (Kennett 2005:230). Establishment of other primary village sites on the
islands appears by the middle Holocene in what would be secondary habitats: those
associated with moderately sized drainages and less access to marine foraging patches. A
process of infilling tertiary habitats on the islands appears to have occurred by the
Middle-Late Period Transition (~1500 BP) when the islands entered the period of highest
population density since their colonization in the Terminal Pleistocene. This was a time
of great social and technological change, when a shell bead currency emerged, and use of
the more seaworthy tomol plank canoe and fishing technologies both increased trade with
the mainland and led to intensive exploitation of offshore fisheries (Arnold 2001; Kennett
2005). Along with resource intensification and increasing diet breadth come signs of
growing status differentiation, increased evidence for interpersonal violence, and
osteological evidence of nutritional stress (Lambert 1994).
Kennett argued that the expression of social conflict, as reflected by lethal and
sublethal violence, as groups colonized the lowest ranked habitats is more consistent with
129
the despotic variant of the IFD (Kennett 2005; Kennett et al. 2009; Winterhalder et al.
2010). Thereby the emergence of social inequality is tied directly to population density,
resource intensification, and technological change at a specific point in cultural history.
The original Channel Islands model has since been further refined with better integration
of ecological, temporal and spatial data in a GIS system (Kennett et al. 2009), and by
incorporating a Bayesian approach to the chronological and geographic sampling that
minimizes the effect of missing data in the record (Winterhalder et al. 2010), indicating
directions for applications in other archaeological settings and geographic scales.
The record of episodic expansion of Polynesian peoples across the Pacific has
also been explored in terms of the IFD (Kennett et al. 2006) and the IDD (Kennett and
Winterhalder 2008), with population pressure, agricultural intensification and ecological
degradation considered as key factors in both triggering pulses of migration and the
emergence of status differentiation in the form of hereditary chiefdoms. As summarized
by Anderson (2001), the initial colonization of Polynesia is signaled by the spread of
Lapita culture into Fiji and West Polynesia between ca. 1300 and 600 BC, which is
considered part of a broader dispersal of speakers of Austronesian languages (Diamond
and Bellwood 2003). Archaeologically, Lapita culture is recognized by the presence of
distinctive dentate-stamped pottery that is distributed into Remote Oceania as far as
Tonga and Samoa, and early footholds on these islands are primarily associated with
coastal rather than interior settlements (Anderson et al 2001; Kirch and Hunt 1988).
Further expansion appears to have stalled for roughly 1600 years before the earliest
documented settlements in East and South Polynesia (AD 1100-1000; e.g., Society
Islands, Marquesas, Hawai’i), with more remote islands such as Rapa Nui (Easter Island),
130
Rapa, and New Zealand settled by about AD 1200 (summarized in Kennett and
Winterhalder 2008).
Kennett et al. (2006) view the hiatus as the result of a period of demographic
infilling in the islands of Fiji and West Polynesia as colonizing populations increased
over time. Evidence for increasing population density is inferred from a range of
archaeological indicators: larger site sizes; decreased residential mobility; settlement
expansion into island interiors; and agricultural intensification indicated by terracing and
irrigation systems (Kennett and Witnerhalder 2008). Considering this process at the scale
of individual islands, this settlement progression is consistent with predictions of the IFD.
In a mixed foraging/agricultural economy, coastal settlements that offer optimal access to
both marine and terrestrial resources would be higher ranked than interior habitats, and
therefore should be occupied first. When population densities increased to the point
where habitat suitability declined for the highest ranked habitats the disadvantages of
interior settlements became less significant, and migration occured. As Kennett et al.
(2006) note, and Kennett and Winterhalder (2008) develop more fully, this process also
likely involved some aspects of despotism as well, pointing to Kirch’s (2000) inference
from linguistic evidence that hierarchical sociopolitical traditions existed among Lapita
groups. Access to the best settlement locations in such a society could be effectively
restricted by certain individuals or groups (and also vigorously contested through intra-
group conflict), leading to a population distribution and land-use pattern more consistent
with the IDD. Another crucial aspect of density-dependent declines in habitat suitability
in the Polynesian case is the environmental consequence of resource intensification that
led both to loss of island flora and fauna targeted by foragers, and increased soil erosion
131
in agricultural contexts. As Lapita populations in-filled the islands, some prime
settlement locations became restricted, and more marginal ones became degraded; as
equilibrium population distributions – free or despotic – were reached after more than a
millennium in West Polynesia. Within this context, another wave of exploration and
migration to new island habitats began. Conceived of in this way, the tempo and mode of
Polynesian expansion can be understood through the integration of ecological, cultural
and ideological factors using a generalizable model that is at once diachronic, spatially
scalable, and open to inclusion of a variety of new archaeological and ecological
observations (Kennett and Winterhalder 2008; Winterhalder and Kennett 2006).
Applying the IFD to Household Settlement at Uxbenká
At Uxbenká in southern Belize the establishment of household settlement groups
should proceed from the highest ranked habitats in the Late Preclassic and Early Classic
into lower ranked habitats as the landscape fills through the Classis period. Because
settlement mobility becomes reduced due to political and social circumscription
throughout the Classic Period, intensive strategies will be employed to offset
climatically-driven and density-dependent habitat degradation around settlement groups.
The model predicts that higher ranked settlement groups will have earlier initial dates of
occupation and longer periods of occupation, and those lower ranked will have later dates
of initial occupation and will have been occupied for a shorter period. Chronological data
to test these predictions are drawn from a combination of archaeological and
chronometric research, including: AMS 14C radiocarbon dated samples recovered from
excavations; Bayesian modeling of selected sequences from Groups A and B (Culleton et
132
al. 2012); temporally diagnostic ceramics recovered from deposits; dedicatory dates on
stelae (Group A); and the presence of architectural features such as ballcourts, which are
typically assigned to the Late Classic Period. These data for the large civic-ceremonial
architectural groups and the domestic settlement groups (SG) are summarized in Table
5.1. Cases where the archaeological data contradict the model expectations will point to
other factors affecting settlement choices that need to be considered, such as the role of
competition and social dominance described by the despotic variant of the IFD (Kennett
et al. 2006, 2008).
Table 5.1. Chronological data on 22 settlement groups (SG) and core groups considered in IFD modeling.
SG
Latest Preclassic (AD 1-300)
Early Classic I
(AD 300-425)
Early Classic II
(AD 425-600) Late Classic (AD 600-800)
1 C C C C
3 R
4 R C C,R C
5 R R
20 C,R
21 R R
23 R
24 R
36 - - - -
38 R R
39 R
50 - - - -
51 - - - -
53 - - - -
54 R
55 - - - -
56 - - - -
57 - - - - Core
Group
A R D,R D,R D,R
B R R R A,C,R
G - - - -
I R R R A Chronological attribution based on : A. architecture (e.g., ballcourt); C: diagnostic ceramics; D: dedicatory date on stela; R: radiocarbon date or modeled event; -: no data for site.
133
Key environmental parameters influencing settlement decisions around Uxbenká
are considered to be agricultural potential, hydrology (i.e., access to freshwater), and
linear distance from the site core as measured from Group A (specifically the peak of
Structure A1). Similar to the approach of Kennett et al. (2009), the selection of these
particular model variables is supported by knowledge of ancient and modern Maya land
use and custom, as well as personal experience on the landscape during several years of
fieldwork. Each variable is discussed below to develop the decision-making context and
provide a rationale for its inclusion in the IFD model. All of the environmental data were
incorporated into a GIS for quantification and analysis along with the settlement survey
data gathered by the Uxbenká Archaeological Project since 2005 (Figure 5.2).
Agricultural Productivity
The ancient Maya inhabitants of Uxbenká, like their contemporary Maya
counterparts in the village of Santa Cruz, were primarily subsistence farmers who relied
heavily upon maize as a staple crop along with secondary crops such as manioc, beans,
squash, and cacao. As such, proximity and access to the most productive lands is
expected to be one of the main criteria for household site selection (or extended
household group). A measure of soil productivity around Uxbenká has been developed
from empirical data on maize yields in the contemporary milpas cleared and planted by
Santa Cruz farmers in 2009 and 2010 (see Chapter IV). Yields from each plot (expressed
as bulk maize yields normalized to account for planting density; kg/ha/planting) were
used to interpolate the productivity across the landscape with the Spatial Analyst toolset
in ArcGIS using a Nearest Neighbor method. The result is a raster surface with a
134
Figure 5.2. Composite showing ecological variables incorporated into the IFD model of settlement at Uxbenká: agricultural productivity (raster gradient); hydrology; and distance from the site core. Periods of earliest occupation for settlement and core groups are noted where data exist (image created by B. Culleton and C. Ebert).
resolution of 18.5 m where each 18.5 x 18.5 m (0.034225 ha) cell has an associated yield
index value ranging from 11.8 – 92.9 kg/ha/suk’ub, with an average value of 47.9
kg/ha/suk’ub (see Chapter IV). There is a clear north to south gradient of greater to lesser
maize yields and this maps on to the known distribution of more prized box lu’um (dark
soils) and less productive chik lu’um (red soils) as described by modern farmers in Santa
Cruz village. The chik lu’um is located primarily between the village and the Rio Blanco.
A 0.5 km-radius catchment was defined around each settlement group and core group
completely within the raster coverage, and the individual yield index value for each cell
135
(~2300 cells within each catchment) was compiled in a database and the average yield
value was used to rank each SG and Core Group in terms of agricultural productivity.
Agricultural productivity is assigned a 50% weighing in the determination of overall rank
due to its perceived importance.
Hydrology
Compared to California’s Northern Channel Islands, access to freshwater in
tropical southern Belize is considered a less crucial but still important factor in settlement
decisions. The presence of the relatively large drainage of Rio Blanco and some of its
main tributary streams would provide access to water even during the depths of the dry
season and the time and effort involved in transporting water during the driest times of
the year would still make locations near larger streams more favorable for settlement, all
other things being equal. To quantify the hydrologic potential of each SG and Core
Group, all of the stream segments in the Uxbenká vicinity were ordered according to the
Strahler’s (1957) method. The locations of these streams (and by extension, what is
defined as a stream) are taken from a digitized and orthorectified hydrologic GIS layer
derived from the 1950s British Ordnance Survey maps for Belize. Stream-ordering is a
convenient approach for characterizing the relative discharge between drainages and
watersheds from essentially analog geographic data, especially in the absence of a higher-
resolution digital elevation model (DEM) from which the areas of watersheds could be
more accurately defined and quantified. The approach is as follows. Any stream in the
broader hydrological system with no tributaries (i.e., those at the headwater of any-sized
drainage) is designated a 1st order stream. Where two 1st order streams join the segment
136
downstream from the confluence is designated a 2nd order stream. When two streams of
differing order meet, the downstream segment remains the higher order of the two.
Where two equally ordered segments meet, the downstream segment is of the next
highest order. For example, if a 1st and 3rd order stream meet, the next segment remains
3rd order; if two 3rd order streams meet the downstream segment is 4th order.
After ranking each segment, the same 0.5-km radius catchment was applied to
each SG and Core Group, and the length of streams of each order was quantified. Though
Strahler’s (1957) method does not perfectly correlate with overall discharge or watershed
area in every case, a hydrological value was devised that weighted stream lengths
geometrically by order to reflect the geometric nature of both hydrological cross-section
and watershed area, and their relationship to discharge. Length of 1st order streams was
weighted at ×1, 2nd order at ×2, and 3rd order at ×4 (i.e., ×20, ×21, ×22) and summed, and
the sites were ranked in terms of hydrology based on this value. It is worth noting that
only two sites, SG 56 and Group I, both to the west of the site core, had a 3rd order stream
within their catchment. Hydrology is given a 30% weighting in the overall rank for each
location.
Distance from the Site Core
Proximity to the site core is a variable that incorporates both social and ecological
aspects of settlement decision-making into the model, and presumes an added resource
potential provided by the urban center of Uxbenká (or any urban center) and what this
offered people in terms of social, commercial, ideological, or subsistence opportunities,
and their desire to be located near them. Some of the attractions an urban center held
137
would likely be greater access to: 1) rarer goods not produced in household economies,
such as salt, cacao, obsidian, or finer pottery in markets; 2) the exchange of information
and maintenance of social ties among commoners; and 3) participation in social and
religious events conducted by elites and specialists (a “theater state”; Demarest
2004:149-160; Zimmerman Holt 2009). On the broader regional scale, closer proximity
to the site core could offer households greater protection from aggression by outside
groups. At the same time, we should keep in mind the possible desire of some individuals
to settle farther from the reach of ruling elites and their ability to extend physical,
economic and social influence over commoners. Proximity to the site core was measured
as the linear distance of each SG or Core Group to the peak of Structure A1, the largest
structure in Group A, which is the location of the earliest known activities at Uxbenká
(Culleton et al. 2012). Sites were then ranked in ascending order according to distance
from Structure A1. The choice of linear distance in this hilly and incised landscape rather
than a least-cost path is justified by a comparison of established farmers’ roads (i.e.,
trails) emanating from nearby Santa Cruz village with a series of least-cost paths
generated using the 30 m-resolution DEM for the area (E. Kalosky, pers. comm., 2010).
The roads, which farmers travel on foot to reach distant milpas (often backing loads in
excess of 50 kg), radiate as nearly linear paths from the village and ignoring slope and
terrain features, contrary to what would be predicted from an slope/elevation derived
least-cost model. Practical experience cutting trails through bush with these farmers also
indicates that most will choose the shortest path in terms of distance rather than the one
with the gentlest slope and I suspect that the same strategy was used by the ancient Maya
138
as they traversed this landscape. Distance from the site core is assigned a 20% weighting
in the overall rank for each location.
The Model
Values and rankings for each of the three variables, and the overall rank for each
SG or Core Group location are presented in Table 5.2. Overall rank is calculated as the
weighted average of each rank where:
Weighted Score = (Productivity Rank × 0.5) + (Hydrology Rank × 0.3) + (Distance Rank × 0.2).
Examples of high-and low-ranked settlement groups are depicted in Figure 5.3.
The highest ranked site location in the available sample is Group A itself, which ranks in
the first quartile for productivity (at 4) and proximity to the site core (at 1), and at the top
of the second quartile for hydrology (at 6). The proximity rank is problematic, since
Group A is the datum from which all the other distances are measured, so obviously it’s
the closest site to itself. Other settlement groups close to the site core are also ranked
relatively high, such as SG 20 (ranked #3), which is located on the ridge between Groups
A and B and contains a late Preclassic deposit buried under a large mound of fill (Prufer
et al. 2011; cf. Chapter III), and SG 21 (ranked #6), which is a small settlement group set
on a finger ridge near Group F. In these two cases proximity to Group A also maps onto
the northerly distribution of highly productive soils and this contributes to their high rank
along with proximity. In contrast SG 5, located immediately to the south of Group A and
ranked 4th in proximity, is on poorer land that is only ranked 16th in this sample (i.e.,
139
Table 5.2. Environmental parameters used to rank settlement groups and core groups at Uxbenká.
Maize
Productivity Hydrology, Length of Ordered Streams Distance
from Group A Overall
SG
Bulk kg/ha/
planting Yield Rank
1st Order
(m)
2nd Order
(m)
3rd Order
(m) Wtd
Value Rank km Rank
Wtd Scor
e Rank
1 63.37 1 536.5 638.3 0.0 1813.0 8 1.73 12 5.3 2
3 40.63 18 649.9 0.0 0.0 649.9 17 2.06 15 17.1 20
4 36.52 19 572.8 0.0 0.0 572.8 18 1.86 13 17.5 22
5 47.08 16 1237.1 0.0 0.0 1237.1 13 0.45 4 12.7 14
20 54.38 5 1731.7 0.0 0.0 1731.7 9 0.38 2 5.6 3
21 50.73 10 1280.1 0.0 0.0 1280.1 11 0.44 3 8.9 6
23 47.47 15 1084.3 840.7 0.0 2765.7 4 0.82 7 10.1 9
24 50.74 9 981.7 518.4 0.0 2018.4 7 1.00 8 8.2 5
36 36.51 20 604.2 1158.9 0.0 2922.1 2 1.06 9 12.4 12
38 34.09 22 625.6 1101.9 0.0 2829.3 3 1.17 10 13.9 15
39 34.73 21 517.4 1046.2 0.0 2609.8 5 1.20 11 14.2 16
50 47.70 14 390.4 0.0 0.0 390.4 22 2.23 18 17.2 21
51 48.09 13 512.9 0.0 0.0 512.9 20 2.41 20 16.5 18
53 52.80 6 1299.1 45.9 0.0 1390.8 10 2.59 21 10.2 10
54 50.01 11 572.7 0.0 0.0 572.7 19 2.36 19 15 17
55 54.61 3 1119.0 0.0 0.0 1119.0 15 2.16 17 9.4 8
56 61.85 2 806.6 774.6 1489.9 8315.4 1 2.82 22 5.7 4
57 51.18 7 500.4 0.0 0.0 500.4 21 1.94 14 12.6 13 Core
Group
A 54.55 4 2047.3 221.2 0.0 2489.7 6 0.00 1 4 1
B 51.01 8 1179.8 0.0 0.0 1179.8 14 0.60 5 9.2 7
G 49.72 12 1276.4 0.0 0.0 1276.4 12 0.73 6 10.8 11
I 43.63 17 385.7 0.0 128.7 900.4 16 2.13 16 16.5 19
close to the bottom of the third quartile). In general, sites to the north are ranked higher
than those to the south reflecting the heavier weighting of productivity in the overall
model.
The predictions of this model are that if IFD conditions prevailed during the
establishment and settlement expansion of Uxbenká, then the earliest settlements should
be found in the highest ranked habitats in terms of agricultural productivity, hydrology
(i.e., access to freshwater), and proximity to the site core. Furthermore, the highest
ranked sites should show more persistent occupation throughout the Classic Period.
140
Figure 5.3. Examples of high and low ranked settlement groups based on the ecological variables within a 0.5 km catchment radius.
Chronological data to test these predictions were summarized above in Table 5.1, and is
integrated in with the habitat rankings in Table 5.3 below to allow for comparisons
between individual variables and composite rankings.
In Table 5.3, a settlement chronology consistent with IFD predictions would be
represented as the earliest and most continuously occupied sites to the left (highest
ranked), and sites occupied later in time to the right (lower-ranked). That is, the dots in
each matrix would tend to fall above and to the left of a diagonal from bottom left to
upper right. The overall picture of habitat suitability proposed here is somewhat
141
Table 5.3. Chronological data and site rankings by individual variables and overall rankings. Ranked by Yield First Quartile Second Quartile Third Quartile Fourth Quartile Late Classic � - - � - - � � � - - - � � � - � � E Classic II � - - � - - � � - - - � � � � - � E Classic I � - - � - - � - - - � � - L Preclassic � - - � � - - � � - - - � � � -
SG or Group 1 56 55 A 20 53 57 B 24 21 54 G 51 50 23 5 I 3 4 36 39 38
Ranked by Hydrology
Late Classic - - � � � � � - � - � - � � � - - -
E Classic II - - � � � � - - � � - � � � - - -
E Classic I - - � � - - � - � � - - -
L Preclassic - - � � � - � - � � - � � - - -
SG or Group 56 36 38 23 39 A 24 1 20 53 21 G 5 B 55 I 3 4 54 51 57 50
Ranked by Dist to Core Late Classic � � � - � - � � � � - � - - � - - - E Classic II � � � - � - � � � - � � - - - - - E Classic I � � - - � � - � - - - - - L Preclassic � � � � � - - � � - � - - - - -
SG or Group A 20 21 5 B G 23 24 36 38 39 1 4 57 3 I 55 50 54 51 53 56
Late Classic Period, AD 600-800; Early Classic Period II, AD 450-600; Early Classic Period I, AD 300-450; L Preclassic, Latest Preclassic Period (AD1-300). �: Evidence for site use; -: no chronological information for the site.
142
Table 5.3. (cont.) Chronological data and site rankings by individual variables and overall rankings. Ranked by Yield and Hydro, Unwtd First Quartile Second Quartile Third Quartile Fourth Quartile
Late Classic - � � - - � � � - - � � - � - � - � E Classic II - � � � - - � - - � - � - � � - � E Classic I - � � - - � - - - - � - � L Preclassic - � � � - - � � - - - � - � - �
SG or Group 56 1 A 20 24 53 55 23 21 B 36 G 38 39 57 5 54 51 I 3 50 4
Overall Rank, Unweighted Late Classic � � � - � � - - � - � - - � � � - -
E Classic II � � � - � - - � � - - - � � � - -
E Classic I � � - � - - - - - � � - - L Preclassic � � � � - � - - � - - - � � - -
SG or Group A 20 1 21 24 56 23 B G 36 5 38 55 39 53 57 54 I 3 4 51 50
Overall Rank, Weighted
Late Classic � � - � � - � - - - - � � � - � - �
E Classic II � � - � � - - - - - � � - � � - �
E Classic I � � - � - - - - - - � - �
L Preclassic � � � - � � - - - - - � - � - �
SG or Group A 1 20 56 24 21 B 55 23 53 G 36 57 5 38 39 54 51 I 3 50 4
Late Classic Period, AD 600-800; Early Classic Period II, AD 450-600; Early Classic Period I, AD 300-450; L Preclassic, Latest Preclassic Period (AD1-300). �: Evidence for site use; -: no chronological information for the site.
143
more mixed, suggesting that unaccounted factors need to be considered. The earliest
occupied sites documented at Uxbenká are the main Groups A, B, D (not ranked here),
and I, and settlement groups 1, 4, 5, 20, 21, all of which indicate occupation during at
least the latest phase of the Late Preclassic, ca. AD 1-300 (Culleton et al. 2012; Prufer et
al. 2011). As noted, several of these are relatively close to the site core (SG 5, 20 and 21),
which would suggest that proximity predicts habitat suitability fairly well. However, SG
1, SG 4 and Group I are located relatively distant from Group A, suggesting that these
earlier sites were selected for reasons other than proximity to Group A during the
Preclassic Period. SG 1 is ranked highest in terms of yield, and in the top quartile in
weighted and unweighted overall ranks, suggesting that the choice to settle there was
guided largely by agricultural concerns rather than association with the core area. Group I
and the nearby SG 4 are located to the west of the site core, and rank in the lowest
quartile overall, largely due to the low ranking in productivity and hydrology, but also
affected by distance from the core. The early settlements near the site core are consistent
with the IFD model, but Group I and SG 4 do not simply conform poorly to the model
predictions, their low rankings contradict the model outright. This leads to several
considerations of the model and the specific nature of Group I as a main architectural
group.
The large, highly visible architectural groups that are concentrated in the core of
Uxbenká (Groups A-G and K) are practically contiguous along two ridgelines within
sight of each other. Groups A and K form the eastern complex, and Groups B-G form the
western complex. Bayesian analysis of the radiocarbon evidence from Groups A, B, and
D indicate that initial clearing and construction at these sites occurred during the latest
144
part of the Late Preclassic. In contrast, Group I appears from the present survey of the
area to be a rather isolated group, but definitely a substantial one with an elite presence
(Reith et al. 2011). AMS dates on early deposits suggest a Late Preclassic date for initial
construction, i.e., probably contemporary with the construction of groups in the core area,
and dates on a tomb containing finely made ceramics vessels, jade beads and earspools
indicate an elite presence there during the Early Classic Period. A ballcourt suggests that
Group I served as a locus of civic and/or ceremonial activities during the Late Classic
period as well (Reith et al. 2011). The picture that emerges from these data is of a
detached center developed during the Late Preclassic by, perhaps, a group competing
with those that established and expanded the main core area of Uxbenká. From this
perspective, Group I’s lower-ranked location would be more consistent with a despotic
distribution, suggesting that individuals involved in settling and elaborating the core
groups prevented these people from establishing themselves in the core during the Late
Preclassic. That is, the early presence of Group I in a relatively marginal habitat is
consistent with the IDD more so that the IFD. Aside from the lack of known stelae at
Group I, similar features of elite expression are found at Groups A, B and I throughout
the Classic Period, suggesting that status rivalry and competition persisted between the
core and this detached faction after the Preclassic Period. However, nothing is known
about the political history of this inferred rivalry, and it is possible that the elites living in
the site core established and maintained hegemony over the Group I faction at various
points during the Classic Period.
If such a rivalry existed including proximity to Group A as a variable in the IFD
model should be reconsidered. If some form of despotic behavior existed during the Late
145
Preclassic and it constrained free settlement then, proximity to a detached center like
Group I might figure more prominently in settlement decisions. Although Group I is less
extensive and elaborate than the groups in the core area, it likely would have offered
similar benefits of social, economic, and ideological interaction for people at settlements
more distal to the site core. Accounting for this sort of “social gravity” in the model could
be done by proposing the presence of Groups, A, B, I and so on, a priori during the Late
Preclassic, and reckoning settlement group distances from the nearest main group, rather
than only Group A. This would help explain the early and persistent occupation of a site
like SG 4, whose proximity to an already established Group I would raise its habitat
suitability. Under this revised model, the early settlement of SG 1 would still be primarily
explained by the highly ranked agricultural potential of its relatively remote location.
It is also possible that other factors influencing site selection need to be
considered to explain the relatively early establishment of Group I and SG 4. The three
factors involved in the proposed IFD model are primarily oriented towards evaluating
habitat suitability with respect to internally-oriented criteria. Maize productivity relates
most directly to the commoner household subsistence economy, and then secondarily
toward the broader polity as surplus maize is given in tribute to elite functionaries or
bartered for other goods. Hydrology, or water availability, is also directly tied to the
concerns of the household economy in terms of labor required to obtain suffiencient
water for daily needs. Distance from the site core, as described above, touches on several
social and economic advantages of access to the concentration of civic and ceremonial
power that made Uxbenká a sociopolitical entity – a small city – in and of itself. Turning
to interactions with people and polities outside of Uxbenká, peripheral outposts might
146
offer individuals the opportunity to observe and mediate access to the site core along
transport routes between other polities. Group I and SG 4 are on either side of the
existing San Antonio-Jalacte Road to the west of Santa Cruz village, and least-cost
models of area suggest that the same route would have been favored for travel to and
from what is now eastern Guatemala (K. Prufer, pers. comm. 2012). The ability to
influence commerce and diplomacy by restricting trade routes is a possible factor that
might raise the habitat suitability of the Group I locality despite generally lower
agricultural productivity. If the area served as one of the entrances to the polity, however,
it seems more likely that the elites at Group I would have been politically integrated with
the elite apparatus in the site core, rather than a competing rival faction. This line of
thinking also raises the question of whether there would have been other potential routes
to access the core area – e.g., across the Rock Patch to the south towards Pusilha, or to
the northeast towards Lubaantun and Nim Lit Punit - and whether there are similar
outposts or garrisons controlling access there as well.
Further Work towards Understanding the Ideal Free and Despotic Settlement
Models for Uxbenká
A focus of ongoing work at Uxbenká is augmenting the chronological records for
many of the sites in the sample. It is clear that Groups A, B, and I have longer records of
occupation than most settlement groups. Much of this is owed to better documentation at
those locations because of the larger effort devoted to their excavation over the years, as
well as the greater frequency of secure contexts within large structures from which to
collect radiocarbon samples to establish absolute chronology (e.g., pit features, plaster
147
floors, rebuilding events, tombs; Culleton et al. 2012; Prufer et al 2011). Smaller
settlement groups with less elaborate architecture and simpler construction histories often
do not provide adequate contexts for sampling because Late and Terminal Classic
deposits, if they exist, are also mixed by bioturbation and other processes into the
present-day A Horizon (Culleton et al. 2012; Webster et al. 2004). There is no way
around this obstacle for radiocarbon dating in many settlement groups, but the problem
may be ameliorated through the ongoing research of ceramic types present in these
deposits. Preliminary work on the ceramic assemblages of SG 1 and SG 4, for example,
documents components attributed to the Late Preclassic Period through the Classic
Period. This is also supported by AMS 14C dates from the site. Further refinement of the
diagnostic ceramic sequence will help flesh out the settlement chronology and provide a
stronger test of the IFD model presented here.
Additional fieldwork is being conducted to expand the database of maize yields
surrounding Uxbenká and Santa Cruz. These new data may alter the rankings for
individual settlement groups, but the general pattern of greater productivity in the north
and lower productivity to the south will likely remain unchanged. The broader areal
coverage will, however, extend the yield raster and allow for additional known settlement
groups with existing chronological data to be included in the sample of sites considered
in this analysis. With a larger sample of sites further complexities in the settlement
history of Uxbenká can be explored within the IFD and IDD models developed here.
148
Conclusions
An IFD/IDD model of habitat suitability and settlement expansion was developed
for the ancient Maya center of Uxbenká and tested against settlement data from the end
of the Late Preclassic Period (AD 1-300) through the Late Classic Period (AD 600-800).
A sample of 22 known civic/ceremonial architectural groups and household settlement
groups was ranked in terms of three variables: agricultural potential, access to potable
water, and proximity to the site core. These variables were quantified from empirical data
on contemporary maize yields in the area, stream ordering, and linear distance from the
core area, incorporated into a GIS database along with archaeological survey coverages
of known settlement sites. These variables were combined into a weighted overall
ranking of habitat suitability for each settlement location in the sample. The prediction of
the IFD model is that the highest ranked habitats should be settled first, and as population
density increases, settlements will expand into less favorable habitats over time.
Comparison of the existing archaeological chronology with settlement ranks
shows a general conformity with the IFD, in that several of the earliest (i.e. Late
Preclassic) settlements are found in high-ranked locations near the site core (e.g., SG 5,
SG 20, and SG 21), and in the most agriculturally productive areas away from the site
core (e.g., SG 1). Two other Late Preclassic settlements – civic-ceremonial Group I and
a smaller household settlement SG 4 – defy the predicted pattern and are found in much
lower-ranked (3rd and 4th quartile) habitats to the west of the Uxbenká’s urban core. The
presence of these sites in marginal habitats early in the settlement history of Uxbenká
may be interpreted as evidence for early despotic behaviors practiced by elites seeking to
exclude certain segments of the population from establishing settlements near the site
149
core. If so, this suggests that competition and status rivalry developed between at least
two competing elite groups: one located in the site core and the other in the detached
center at Group I. Alternatively, Group I may have been positioned to mediate access to
the site core from travellers outside of the polity, and functioned as a garrison or outpost.
In that case, the Group I population was more likely to have been politically integrated
with core elites, rather than a competing rival faction. Further work is needed to improve
the archaeological chronology and incorporate more sites into the analysis, but the results
demonstrate the utility of formal IFD and IDD models for exploring the ecological and
social factors affecting population distributions in the past and for identifying and
explaining instances of status competition in the archaeological record.
150
CHAPTER VI
CONCLUSIONS AND PROSPECTS
The archaeological research presented in this dissertation is the outcome of
several years of collaborative work with colleagues in the field and lab with the aim of
understanding the connections between land use, ecology and settlement at the ancient
Maya center of Uxbenká, Belize. It is the result of the kind of interdisciplinary effort that
marks the higher ambitions of archaeological research, which, as van der Leeuw and
Redman (2002) argue, is to place “archaeology at the center of socionatural studies.”
Doing so means attempting to bridge methodological, theoretical and cultural gaps
between disciplines, and a willingness to share our data, expertise, and to help shoulder
the burdens of interpretation and analysis. My work at Uxbenká is a small contribution to
the larger on-going research project there, and to Maya archaeology in general, but
several of the approaches outlined here may show promise for broader application as
collaboration continues.
The Bayesian chronology developed here provides new insights into the
developmental history of Uxbenká’s urban core and provides a statistical framework for
future chronological refinement. The earliest leveling and clearing at Group A (the Stela
Plaza) began during the Late Preclassic at cal 50 BC – AD 220, roughly 100-200 years
earlier than previously thought (Prufer et al. 2011). This was followed by similar
landscape modifications at Group D (cal AD 20-240) and Group B (cal AD 60-310) and a
period of multiple plastering and remodeling episodes in both plazas. The leveling and
construction during the Late Preclassic and the Early Classic that established the nascent
151
urban core of Uxbenká preceded all evidence for dated stone monuments at the site, as
the earliest known stela was dedicated in AD 378. Based on the available evidence there
is relatively little construction in the site core that dates after the Early Classic Period
from ca. AD 400-600. However, the Group A plaza was substantially replastered in the
Late Classic at cal AD 550-770 along with the construction and dedication of a staircase
in Group B (Structure B1; cal AD 650-770). These events coincide with the dedication of
stela at Uxbenká and the appearance or expansion of other regional polities (e.g., Pusilhá,
Lubaantun, Nim Li Punit) that is possibly tied to increased interaction with the Petén
region in Northern Guatemala (e.g., Tikal). Secure Terminal Classic contexts have been
difficult to identify, but remain a focus of ongoing investigations at Uxbenká.
The geoarchaeological work at Uxbenká has defined two episodes of cultural
activity that precede the earliest evidence for the leveling and construction of buildings in
the urban core. Non-diagnostic ceramic sherds recovered from these A horizons provide
the earliest evidence for human occupation in what later became the urban center. This is
currently the earliest evidence for human activity in the area and is consistent with the
hypothesis that a small farming population first colonized the area between ~900 and 800
BC. This pioneering agricultural activity also occurred during a dry climatic interval that
may have destabilized the landscape further. Soil stability during the Middle Preclassic
(~770-520 cal BC) occurred during a drying trend that was punctuated by several severe
dry periods. This suggests that the landscape is fairly resilient under naturally dry
conditions. Destabilization again coincided with the appearance of pottery and stone tools
in the sediments at ~300 cal BC, but also with one of the more severe drying trends that
likely contributed to deforestation and erosion. I argue that the absence of agricultural
152
terraces and other soil retention features in the area surrounding the urban core results
from naturally occurring soil retention features and the rapid decomposition of the
mudstone bedrock favoring soil replenishment. I also argue that the overall stability of
the landscape in the urban core between ~60 BC and AD 900 resulted from the absence
or reduction of swidden cultivation in what was essentially an urbanized landscape used
for civic-ceremonial activities and possibly stabilized by urban gardens and the
cultivation of economically valuable tree crops. An episode of mass-wasting in the urban
core occurred during the Early Classic sometime between cal AD 280-610, and is
attributed to possible tectonic activity and associated hillslope failure, rather than human
activities in the site core. Increased erosion and the burial of the Late Classic Period
landscape is coincident with increasing evidence for swidden agriculture in the site core,
possibly by a remnant or returning population of farmers after the political collapse of
Uxbenká that occurred in the context of climatic and social instability during the
Terminal Classic Period.
The results of the geoarchaeological work suggest further avenues to explore. The
presence of pottery in the early paleosols is currently the earliest evidence for ceramics in
southern Belize. They are Middle Preclassic in age and this is consistent with the
relatively late adoption of ceramics elsewhere in the eastern Maya Lowlands. The lack of
diagnostic slip or discernable vessel form leaves these sherds as tantalizing evidence of a
human presence, but with no indication of cultural or geographic origin. Thin-section
studies and element analysis of the ceramic paste holds the possibility of identifying a
local or exotic origin for the pieces, and might allow their age to be confirmed by
comparison to better preserved specimens from areas such as the Petén or the Belize
153
River Valley. Also the interpretation of changing land use in the site core, from
agricultural to urban during the Classic Period, and returning to swidden cultivation from
in the Terminal Classic might be tested through palynological and paleobotanical studies
on sediments recovered from the paleosol sequences. Shifting land use in the site core
should be identifiable by changing abundances of arboreal and disturbance taxa, and by
the presence or absence of economic cultivars throughout the sequence. Finally, the
development of a high precision speleothem precipitation record from Yok Balum Cave
in the karst ridge roughly 1.5 km south of Uxbenká may clarify the relationship between
the climate change and landscape stability that is somewhat obscured by the
contradictions in the three existing climate records considered in this study.
Contemporary Maya subsistence practices and maize productivity in the area were
used to estimate maximum population potential for the ancient Maya center of Uxbenká.
Maize yields in milpas planted by farmers around the village of Santa Cruz were
quantified during the 2009 and 2010 harvest seasons, and compared with environmental
variables including soil nutrients (e.g., N, P, K, pH, organic and inorganic carbon) and
landscape attributes (e.g., slope, aspect, distance from the village). Maize yields were
found not to correlate with measured variables, with the exception of a very weak
positive correlation with distance from the site core. Planting density, which varies with
the type of maize planted, was found to heavily influence yields and is dependent upon
intercropping with other cultivars and the presence of physical obstacles in cleared
milpas. The lack of correlation between yields and a range of environmental variables is
consistent with other ethnographic studies on maize production that suggest a range
154
confounding factors of soil, weather, maize variety, pests, and farming experience that
ultimately dictate the outcome at harvest.
Taking the average maize yield per area and assuming daily caloric needs for
ancient inhabitants, the maximum sustainable population of the Uxbenká polity during
the Classic Period is estimated to be between 7500 and 13,000 people within a 6 km
radius. This population is modeled at a five-year fallow period, just on the cusp of a short
fallow system suggestive of low level agricultural intensification. The lack of
archaeological evidence for intensive farming strategies (e.g., terracing, field
demarcation, irrigations systems) in the vicinity of Uxbenká is consistent with this model
result. Assuming the elite population resided in the urban core of the site and that it was
5% of the total population, the model predicts the presence of ~525 elites.
A productivity-derived prediction of household density within the ancient
Uxbenká polity provides expectations that can be tested with future archeological work.
The factors affecting maize yields will continue to be investigated by a new cohort of
anthropologists working with farmers in Santa Cruz village. Further directions for
research into past population size include incorporation of more realistic demographic
profiles for the ancient population (i.e., accounting for the distribution of the age and sex
classes of the modeled population with life tables), and the development of more complex
computational models of demographic change over centuries and millennia of land use
and social change.
An IFD/IDD model of habitat suitability and settlement expansion was developed
for the ancient Maya center of Uxbenká and tested against settlement data from the end
of the Late Preclassic Period (AD 1-300) through the Late Classic Period (AD 600-800).
155
A sample of 22 known civic/ceremonial architectural groups (n=4) and household
settlement groups (n=18) was ranked in terms of three variables: agricultural potential,
access to potable water, and proximity to the site core. These variables were quantified
from empirical data on contemporary maize yields in the area, stream ordering, and linear
distance incorporated into a GIS database along with archaeological survey coverages of
known settlement sites. These variables were combined into a weighted overall ranking
of habitat suitability for each settlement location in the sample. The prediction of the IFD
model is that the highest ranked habitats should be settled first, and as population density
increases, settlements will expand into less favorable habitats over time.
Comparison of the existing archaeological chronology with settlement ranks
shows a general conformity with the IFD, in that several of the earliest (i.e. Late
Preclassic) settlements are found in high-ranked locations near the site core (e.g., SG 5,
SG 20, and SG 21), and in the most agriculturally productive areas away from the site
core (e.g., SG 1). Two other Late Preclassic settlements – a civic-ceremonial group
(Group I) and a smaller household settlement (SG 4) – defy the predicted pattern and are
found in much lower-ranked (3rd and 4th quartile) habitats to the west of Uxbenká’s urban
core. The presence of these sites in marginal habitats early in the settlement history of
Uxbenká may be interpreted as evidence for early despotic behaviors practiced by elites
seeking to exclude certain segments of the population from establishing settlements near
the site core. If so, this suggests that competition and status rivalry developed between at
least two competing elite groups, one located in the site core and the other in the detached
center at Group I. However, it is also possible that peripheral settlements that exhibit less
favorable habitat suitablity rankings in the proposed model may offer other advantages in
156
the broader sociopolitical context that have not been accounted for. Peripheral settlements
located at points of strategic access to the site core may have served as points from which
to observe and mediate travel and trade between other regional polities, possibly serving
as garrisons or checkpoints. In such a scenario, the existence of a relatively substantial
elite presence at Group I may be interpreted as an extension of elite political control
throughout the Uxbenká area rather than the center of a competing rival faction.
Further work is needed to improve the archaeological chronology of the
settlement groups around Uxbenká so that a larger sample of sites in a broader range of
habitat types can be incorporated into the analysis to test model predictions. Further
methodological refinements would include the use of Bayesian sampling techniques to
develop finer chronological resolution in the order of settlement expansion, and also to
account for the effects of incomplete settlement survey coverages. Even so, the results of
a relatively simple model formulation demonstrate the utility of formal IFD and IDD
models for exploring the ecological and social factors affecting population distributions
in the past and for identifying and explaining possible instances of status competition in
the archaeological record. A broader application of ecologically-based formal models
holds promise for addressing questions of ancient Maya human-environment interactions
over a range of temporal and spatial scales.
Broader Relevance to Lowland Maya Archaeology
The work presented here on the archaeology of land use at Uxbenká is
fundamentally aligned with the research tradition of cultural ecology in the Maya
Lowlands, while also incorporating more recent theoretical developments in Human
157
Behavioral Ecology (HBE). As outlined by Demarest (2003:22-24), cultural ecology and
economic approaches to understanding Maya culture history, and Mesoamerican
archaeology in general, were widely adopted in the 1960s and influenced research
designs and objectives heavily into the 1970s and early 1980s. The emphasis on
ecological constraints on cultural adaptations as well as attempts to employ hypothesis
testing in research agendas characterized much of Mayanist archaeology during those
decades, leading to an expansion of data-driven empirical work on settlement patterns,
paleodemography, and food production systems at Lowland Maya sites. By the late 1980s
critiques of cultural ecology as being overly deterministic in explanatory power, and
perceived inability to address or explain apparently non-ecological features of ancient
Maya society, such as the ceremonial-religious apparatus of Maya rulers and their elites,
gained ground as an element of the broader post-processual backlash within Americanist
archaeology. The desire to understand the political and social aspects of Maya society
that at first glance are less empirically tractable - but clearly crucial to explaining the
emergence, maintenance and eventual decline of Classic Period Maya polities – drove
research into the arena of political economy of theather states (Demarest 2004; Masson
and Freidel 2002).
As ecological approaches to Mayanist archaeology were gradually being de-
emphasized from the early 1990s on, advances in climate science led to more precise
climate records (primarily lake cores) recovered from Central America. The role of
climate change, specifically drought, in the decline of Maya civilization came to the fore
again (e.g., Hodell et al. 1995), bouyed by the increasing concern about the social
consequences of environmental change among natural and social scientists, political
158
entities, funding agencies, and the general public. This shift back towards ecological
explanation continues to the present, but poses a challenge for archaeologists to
collaborate effectively with climate scientists that desire their work to have broader social
relevance, but may not (yet) be well versed in anthropological theories of societal change.
Much of my research at Uxbenká has attempted to develop the site’s temporal and
ecological context to bridge the gap between archaeological and environmental histories,
so that the effects of human land use and ecological change can be better understood.
Working in collaboration with members of the Uxbenká Archaeological Project and
Maya Socioeconomic Dynamics project, I have helped build the Uxbenká site chronology
using Bayesian techniques, studied the unique geoarchaeological setting of the area,
investigated contemproary maize yields and their implications for past population and
land use, and incorporated these data into a preliminary settlement decision model using
concepts of the Ideal Free and Despotic Distributions from population ecology. This
work complements and augments the more strictly archaeological and ecological work
being conducted by the broader research teams. The approach is not new to Mayanist
archaeology, but can be seen as part of the growing return of ecologically oriented
research of past decades into contemporary research agendas, while employing new
analytical techniques to the study of human environment interactions.
As climate records gain resolution through advances in chronology and sampling
techniques, periods of rapid climate change come into focus and understanding human
responses to them have become more pressing topics of study. Improving the
chronological resolution of Maya culture histories is key if they are to be comparable to
newer climate records so that cause and effect relationships between environmental and
159
cultural change may be properly understood. The use of Bayesian chronology building at
Uxbenká is one example of how existing chronometric and archaeological data can be
integrated to improve site chronologies, and these may be more widely applied
throughout the Maya Lowlands. While ceramic seriation and epigraphic texts have
formed the backbone of Maya site chronologies for decades, better integration with high
resolution AMS 14C dating in a Bayesian framework may yield much tighter absolute
chronologies that are required to test hypotheses of climate-driven social change
throughout the Preclassic and Classic Periods. Given the large body of existing
chronometric data from these various sources at hundreds of Maya centers, there is great
potential to re-evaluate Maya culture history using Bayesian analysis.
The geoarchaeological work at Uxbenká has demonstrated the importance of site-
specific geology and soil formation processes, as well as the value of conducting off-site
investigations. When discussing the ecological setting of the Maya Lowlands, reference
to the limitations of thin limestone soils for maize farmers is extremely common, and of
course it is a broadly accurate description of much of the Maya Lowlands. However, the
mudstone- and sandstone-dominated Toledo Beds on which Uxbenká sits produce
relatively thick soils, and appear to be fairly resilient in the face of swidden agriculture.
That, coupled with the presence of deep joints and fissures in bedrock that act as soil-
retaining structures, appears to have obviated the need for heavy investment in
constructed soil management features during the site’s history. Without excavating
trenches in areas away from the main architectural groups at Uxbenká, the special nature
of the local soils and geology were understood in a way that would have remained
unknown. Further, the record of geomorphic stability and instability in response to land
160
use and climate change was produced in an area where other local proxies of land use –
specifically lake sediment cores – are unavailable or poorly resolved. Off-site work is
extremely valuable for providing proximate records of landscape response when making
comparisons to environmental records derived from other parts of Mesoamerica or
further abroad. Humans respond to the local effects of global environmental change, and
local proxies serve as a test for hypotheses of social change derived from more distal or
regional records.
Interest in estimating ancient Maya population sizes has waned since the early
1990s as a result of the shift away from the larger settlement surveys required when
estimating population from known structural remains, itself part of the general decline of
ecological approaches to Mayanist archaeology. In so far as estimates of population size
and density directly relate to questions about capacity for food production and level of
agricultural intensification, they are crucial for developing the context in which climate
change (e.g., periods of drought) could have altered the economic basis of ancient Maya
societies. The fact that the lands around Uxbenká are currently being farmed by the
modern Maya community of Santa Cruz offered the opportunity to gauge the productive
capacity of the land in a non-mechanized swidden farming system today, and from that
develop maximum population estimates for the ancient polity. The results provide further
predictions about potential settlement densities that can be tested in the course of the
ongoing settlement survey and excavation work by the Uxbenká Archaeological Project.
By estimating possible fallow periods at the site’s peak during the Classic Period, I
suggest that the population may have been just on the cusp of needing to shift to more
intensive agricultural practices, and were engaging in a level of intensification that would
161
typically leave very little archaeological trace. This opens the question of how intensive
Maya agriculture was at any given polity that lacks obvious signs of constructed terraces,
raised fields, and similar adaptations. The lack of such features does not necessarily
equate to extensive land use, but indicates the spectrum of intensification possible, and
the margin of land use flexibility and adaptability inherent in the Maya farming system to
cope with human-induced and external environmental change. Likely many secondary
polities without elaborate soil management structures were indeed still making land use
decisions within that archaeologically obscure margin of intensification.
The application of HBE models to land use and settlement decisions in the Maya
Lowlands is a new approach to understanding land use decisions and dynamics of social
and ecological change. Viewed as an outgrowth of earlier Cultural Ecology paradigms it
offers explicit connections between the socioeconomic context in which individuals
operated and allows predictions of the outcomes of their decision-making process for the
archaeological record. By focusing on individuals, HBE models have the potential for
addressing issues of agency among the commoners that comprised the bulk of ancient
Maya society. Further, by emphasizing the ecological context of human decision making
the integration of HBE into Mayanist archaeology has the potential to bridge
archaeological and anthropological data and theories with those of other natural sciences
including tropical ecology, hydrology, geomorphology, and climate systems. In this sense
HBE may serve as a crucial tool for the broader interdisciplinary endeavors that are
required to address current problems of human responses to enivronmental change by
providing a mutually intelligible framework for communication between disparate
aspects of collaborative projects. The ability to accomodate aspects of climate change
162
with the complexities of human social behavior - particularly in the case of the ancient
Maya where cultural features of religion, ceremony and statecraft may seem ecologically
intractable and therefore inexplicable – is key to understanding the dynamics of the
emergence, maintenance and dissolution of ancient Maya sociopoltical sytems. HBE
offers the potential for broader application and multiple spatial and temporal scales and
should be a productive vehicle for future work in the Maya Lowlands.
The Archaeology of Uxbenká and the Community of Santa Cruz
My work at Uxbenká and the surrounding lands has been conducted with the
permission and and assistance of the Maya community of Santa Cruz, on whose land the
ruins of the ancient polity are found. The Uxbenká Archaeological Project has developed
an excellent working relationship with community members, and helped develop a
community-based organization for the management of cultural tourism related to the
lands and the ruins of Uxbenká, the Uchbenkah K’in Ajaw Association. Because of the
close collaboration with the community members I have been mindful of what
contribution my work could make to the people of Santa Cruz, beyond providing the
short-term economic benefits of wage labor during surveys and excavation. It is a vexing
problem in any circumstance to argue for the practical benefit of archaeological
knowledge for society as a whole, but more so when those benefits to society may appear
to be largely abstract and refer to Euro-American Enlightment goals rather than practical
applications for indigenous farmers.
The men that did the bulk of the physical work excavating with me at Uxbenká
are all farmers engaged in subsistence and cash cropping, and we spent a great deal of
163
time talking about farming and soils. The aspects of my work that deal with soils and
maize productivity probably have the most direct relevance to their concerns as farmers
and householders. One issue that they confront is the long-term sustainability of shifting
swidden agriculture on their communal lands, which they won the right to in the Supreme
Court of Belize in 2007. The geoarchaeology of Uxbenká suggests that since at least the
Middle Preclassic Period, people have engaged in a form of swidden agriculture at
various times and with varying intensity, and that the land responded with periods of
erosion and stability depending on climatic conditions and local land use decisions. This
record exists because the character of the local bedrock provides what are essentially
sediment traps that retain large volumes of soil that today must contribute to the overall
productivity and resilience of the soil to swidden farming and the effects of erosion and
slopewash. In addition, when the mudstone and sandstone bedrock is exposed by forest
clearing it quickly breaks down to form new soil, so that topsoil is relatively rapidly
replenished. Together these aspects of the Santa Cruz lands lend themselves to what
Hartshorn et al. (1984) referred to as the “paradoxical fertility” of the Toledo Beds.
There is not yet enough data to argue that this resilience will persist indefinitely
under current land use practices, which would be one way of defining sustainability.
Longer term study of the local ecology, fertility, dynamics of forest succession and
recovery will be needed to make this argument. However, I would note that the continued
ability to grow maize, rice and other crops on Santa Cruz lands without the emergence of
a grassy wasteland, as feared by Wilk (1991) in the region, is suggestive. Wright et al.’s
(1959) land use recommendation for the area was to log the remaining stands of forest,
clear the bush for pasture and graze cattle for several years, and then devote the land to
164
tree crops such as citrus. He argued that there was very little potential for other economic
uses. By the time of Wright’s contribution to the Hartshorn et al. (1984) field study, he
apparently recognized the aforementioned paradoxical fertility of the region. However,
almost 30 years later it appears that the resilience of the area’s soil to current land use
practices continues. Again, this is attributed to a balance between the advantages of the
local geology and the nature of the communal land use practices, and can’t be said to
apply to every part of southern Belize. However, the geoarchaeological evidence suggests
the capacity of the land to continue to support swidden farming around Santa Cruz with
communal decision-making regarding land use practices. The ongoing work of
archaeologists, ethnographers and ecologists in the Santa Cruz community can draw upon
the geoarchaeological presented here as a baseline for comparing the effects of modern
land use over the longer term. In a small way, perhaps this work can also offer the
community a sense of their place in the longer historical legacy of peoples that have been
making a living farming the land over the last 3000 years in southern Belize.
165
REFERENCES CITED
Abramiuk, M.A., Meurer, W.P., 2006. A Preliminary Geoarchaeological Investigation of Ground Stone Tools in and around the Maya Mountains, Toledo District, Belize. Latin American Antiquity 17:335-354.
Abrams, E.M., Rue, D.J., 1988. The Causes and Consequences of Deforestation among
the Prehistoric Maya. Human Ecology 16:377-395. Adams, R.E.W., 1977. The Origins of Maya Civilization. University of New Mexico
Press, Albuquerque. Adams, R.E.W., 1980. Swamps, Canals, and the Location of Ancient Maya Cities.
Antiquity 54:206-214. Adams, R.E.W., Brown, W.E., Jr., Culbert, T.P., 1981. Radar Mapping, Archaeology,
and Ancient Maya Land Use. Science 213:1457-1463. Aimers, J., Hodell, D. 2011. Societal Collapse: Drought and the Maya. Nature 479:44–
45. Anderson, A.J., 2001. Mobility Models of Lapita Migration. In The Archaeology of
Lapita Dispersal in Oceania: Papers from the Fourth Lapita Conference, June 2000, Canberra (Terra Australis 17). G. Clark, A.J. Anderson, T. Vunidilo (eds.), pp. 15-23. Australia National University Press, Canberra.
Anderson, A.J., Bedford, S., Clark, G.S., Lilley, I., Sand, C., Summerhates, G., Torrance,
R., 2001. An Inventory of Lapita Sites Containing Dentate-Stamped Pottery. In The Archaeology of Lapita Dispersal in Oceania: Papers from the Fourth Lapita Conference, June 2000, Canberra (Terra Australis 17). G. Clark, A.J. Anderson, T. Vunidilo (eds.), pp. 1-13. Australia National University Press, Canberra.
Anselmetti, F.S., Hodell, D.A., Ariztegui, D., Brenner, M., Rosenmeier, M.F., 2007.
Quantification of Soil Erosion Rates Related to Ancient Maya Deforestation. Geology 35:915–918.
Arnold, J.E., 1992. Complex Hunter-Gatherer-Fishers of Prehistoric California: Chiefs,
Specialists, and Maritime Adaptations of the Channel Islands. American Antiquity 57:60-84.
Arnold, J.E., (ed.), 2001. The Origins of a Pacific Coast Chiefdom: The Chumash of the
Channel Islands. University of Utah Press, Salt Lake City. Ashmore, W., 1990. Ode to a Dragline: Demographic Reconstructions at Classic
Quirigua. In Precolumbian Population History in the Maya Lowlands. T.P. Culbert & D.S. Rice (eds.), pp. 63-83. University of New Mexico Press, Albuquerque.
166
Awe, J.J., 1992. Dawn in the Land between the Rivers: Formative Occupation at Cahal
Pech, Belize and Its Implications to Preclassic Developments in the Central Maya Lowlands. Unpublished Doctoral Dissertation, Institute of Archaeology. University College, London.
Barker, G., 2008. The Agricultural Revolution in Prehistory: Why Did Foragers Become
Farmers? Oxford University Press, London. Barlow, K.R., 2002. Predicting Maize Agriculture among the Fremont: An Economic
Comparison of Farming and Foraging in the American Southwest. American Antiquity 67:65-88.
Bateson, J.H., Hall, I.H.S., 1977. The Geology of the Maya Mountains, Belize. Institute
of Geological Sciences, Overseas Memoir 3. Oxford University Press, London. Bayliss, A., Bronk Ramsey, C., 2004. Pragmatic Bayesians: a Decade of Integrating
Radiocarbon Dates into Chronological Models. In Tools for Constructing Chronologies: Crossing Disciplinary Boundaries. Lecture Notes in Statistics 177 (edited by C. E. Buck and A. R. Millard). p. 25-41. Springer.
Bayliss, A., Bronk Ramsey, C., van der Plicht, J., Whittle, A., 2007. Bradshaw and
Bayes: Toward a Timetable for the Neolithic. Cambridge Archaeological Journal 17:1-28.
Bayliss, A., Groves, C., McCormac, G., Baillie, M., Brown, D., Brennand, M., 1999.
Precise dating of the Norfolk timber circle. Nature 402:479. Beach, T. ,1998. Soil Catenas, Tropical Deforestation, and Ancient and Contemporary
Soil Erosion in the Petén, Guatemala. Physical Geography 19:378-405. Beach, T., Dunning, N., Luzzadder-Beach, S., Cook, D.E., Lohse, J., 2006. Impacts of the
Ancient Maya on Soils and Soil Erosion in the Central Maya Lowlands. Catena 65:166-178.
Beach, T., Luzzadder-Beach, S., Dunning, N., Cook, D.E., 2008. Human and Natural
Impacts on Fluvial and Karst Depressions of the Maya Lowlands. Geomorphology 101:308–331.
Beach, T., Luzzadder-Beach, S., Dunning, N., Jones, J., Lohse, J., Guderjan, T., Bozarth,
S., Millspaugh, S., Bhattacharya, T., 2009. A review of human and natural changes in Maya Lowland wetlands over the Holocene. Quaternary Science Reviews 28:1710-1724.
Bellwood, P., 2005. The First Farmers: The Origins of Agricultural Societies. Blackwell
Publishing, Oxford.
167
Bill, C.R., Braswell, G.E., 2005. Life at the Crossroads: New Data from Pusilhá, Belize. Research Reports in Belizean Archaeology 2:301-312.
Binford, M.W., Brenner, M., Whitmore, T.J., Higuera-Gundy, A., Deevey, E.S., Leyden,
B.W., 1987. Ecosystems, Paleoecology, and Human Disturbance in Subtropical and Tropical America. Quaternary Science Reviews 6:115-128.
Birkeland, P., 1999. Soils and Geomorphology. Oxford University Press, New York. Boserup, E., 1965. Conditions of Agricultural Growth: The Economics of Agrarian
Change Under Population Pressure. Aldine Publishing Company, Chicago. Brady, J.E. 1989. Investigation of Maya Ritual Cave Use with Special Reference to Naj
Tunich, Petén, Guatemala. Unpublished Doctoral Dissertation, Department of Anthropology. University of California, Los Angeles.
Braswell, G.E., 1992. Obsidian Hydration Dating, the Coner Phase, and Revisionist
Chronology at Copán, Honduras. Latin American Antiquity 3:130-147. Braswell, G.E., 1996. Obsidian-Hydration Dating. In The Oxford Companion to
Archaeology. B. Fagan (ed.), p. 537. Oxford University Press, New York. Braswell, G.E. (ed.), 2003. The Maya and Teotihuacan: Reinterpreting Early Classic
Interaction. University of Texas Press, Austin. Braswell, G.E., Prager, C.M., Bill, C.R., Schwake, S.A., 2004. The rise of secondary
states in the southeastern periphery of the Maya world: a report on recent archaeological and epigraphic research at Pusilja, Belize. Ancient Mesoamerica 15:219-233
Braswell, G.E., Prufer, K.M., 2009. Political Organization and Interaction in Southern
Belize, Research Reports in Belizean Archaeology 6:43-55. Brenner, M., Rosenmeier, M.F., Hodell, D.A., Curtis, J.H., 2002. Paleolimnology of the
Maya Lowlands. Ancient Mesoamerica 13:141-157. Bronk Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy: the OxCal
program. Radiocarbon, 37:461-74. Bronk Ramsey, C., 2000. Comment on 'The Use of Bayesian Statistics for 14C Dates of
Chronologically Ordered Samples: A Critical Analysis. Radiocarbon 42:199-202. Bronk Ramsey, C., 2001. Development of the Radiocarbon Program OxCal.
Radiocarbon, 43:355-363. Bronk Ramsey, C., 2005. OxCal v. 3.10.
http://c14.arch.ox.ac.uk/embed.php?File=oxcal.html.
168
Bronk Ramsey, C., 2009. Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51:337-360.
Bronk Ramsey, C., Dee, M.W., Rowland, J.M., Higham, T.F.G., Harris, S.A., Brock, F.,
Quiles, A., Wild, E.M., Marcus, E.S., Shortland, A.J., 2010. Radiocarbon-Based Chronology for Dynastic Egypt. Science 328:1554-1557.
Brumfiel, E.,1992. Distinguished Lecture in Archaeology: Breaking and Entering the
Ecosystem – Gender, Class and Faction Steal the Show. American Anthropologist 94:551-567.
Buchanan, B., Collard, M., Edinborough, K., 2008 Paleoindian Demography and the
Extraterrestrial Impact Hypothesis. Proceedings of the National Academy of Sciences USA 105:11651–11654.
Buck, C.E., 2004. Bayesian Chronological Data Interpretation: Where Now? In Tools for
Constructing Chronologies: Crossing Disciplinary Boundaries. Lecture Notes in Statistics 177. C. E. Buck & A. R. Millard (eds.), pp. 1-24. Springer.
Burney, D.A., Flannery, T.F., 2005. Fifty Millennia of Catastrophic Extinctions after
Human Contact. TREE 20:395-401. Cancian, F., 1965. Economics and Prestige in a Maya Community: The Religious Cargo
System in Zinacantan. Stanford University Press, Stanford, California. Carneiro, R.L., 1970. A Theory of the Origin of the State. Science 169:733-738. Carter, W., 1969. New Lands and Old Traditions: Kekchi Cultivators in the Guatemalan
Lowlands. University of Florida Press, Gainsville. Chase, A.F., 1990. Maya Archaeology and Population Estimates in the Tayasal-
Paxcaman Zone, Petén, Guatemala. In Precolumbian Population History in the Maya Lowlands. T.P. Culbert & D.S. Rice (eds.), pp. 149-166. University of New Mexico Press, Albuquerque.
Chase, A.F., Chase, D.Z., 1998. Late Classic Maya Political Structure, Polity Size, and
Warfare Arenas. In A. Ciudad Ruiz, M.Y. Fernández Marquínez, J.M. García Campillo, M.J. Iglesias Ponce de León, A.L. García-Gallo, & L.T. Sanz Castro (eds.), Anatomía de una civilización: Aproximaciones interdisciplinarias a la cultura Maya (pp.1-29). Sociedad Española de Estudios Maya: Madrid.
Chase, A.F., Chase, D.Z., Weishampel, J.F., Drake, J.B., Shrestha, R.L., Slatton, K.C.,
Awe, J.J., Carter, W.E., 2011. Airborne LiDAR, Archaeology, and the Ancient Maya Landscape at Caracol, Belize. Journal of Archaeological Science 38:387-398.
Childe, V.G., 1928. The Most Ancient East. Trubner and Co., London.
169
Childe, V.G., 1951. Social Evolution. Watts, London. Clark, J.E., Cheetham, D., 2002. Mesoamerica's Tribal Foundations. In The Archaeology
of Tribal Societies. W.A. Parkinson (ed.), pp. 278-339. International Monographs in Prehistory, Archaeology Series 15. University of Michigan, Ann Arbor.
Cohen, M.N. 1977. The Food Crisis in Prehistory: Overpopulation and the Origins of
Agriculture. Yale University Press, New Haven, CT. Cowgill, G.L. 1975a. Population Pressure as a Non-Explanation. In Population Studies in
Archaeology and Biological Anthropology: A Symposium. A. C. Swedlund (ed.), pp. 127-131. Memoir 30, Society for American Archaeology.
Cowgill, G.L. 1975b. On Causes and Consequences of Ancient and Modern Population
Changes. American Anthropologist 77:505-525. Culbert, T.P., Kosakowsky, L.J., Fry, R., Haviland, W.A., 1990. The Population of Tikal,
Guatemala. In Precolumbian Population History in the Maya Lowlands. T.P. Culbert & D.S. Rice (eds.), pp: 103-122. University of New Mexico Press, Albuquerque.
Culbert, T.P., Rice, D.S., (eds.), 1990. Precolumbian Population History in the Maya
Lowlands. University of New Mexico Press, Albuquerque. Culleton, B.J., 2008. Potential Soil and Water Conservation Features at Uxbenká. In
Uxbenká Archaeological Project: Report of the 2007 Field Season in the Toledo District, Belize, Central America. K.M. Prufer (ed.), pp. 54-58. Unpublished manuscript. Belize Institute of Archaeology; Belize Institute of Social and Cultural Research; National Institute of Culture and History. Belmopan, Belize.
Culleton, B.J., 2009. Geoarchaeological Investigation and Chronology of Buried Soils
and Cultural Features in the Site Core (SubOps 08-1 and 08-3) at Uxbenká. In Uxbenká Archaeological Project: Report of the 2008 Field Season in the Toledo District, Belize, Central America. K.M. Prufer (ed.), pp. 86-104. Unpublished manuscript. Belize Institute of Archaeology; Belize Institute of Social and Cultural Research; National Institute of Culture and History. Belmopan, Belize.
Culleton, B.J., 2010. Geoarchaeological Investigation and Chronology of Buried Soils
and Cultural Features in the Core and Settlement Areas around Uxbenká (SubOps 09-1, 09-2, 09-6 & 09-22). In K.M. Prufer (Ed.), Uxbenká Archaeological Project: Report of the 2009 field season in the Toledo District, Belize, Central America (pp. 43-63). Unpublished manuscript. Belmopan, Belize: Belize Institute of Archaeology; Belize Institute of Social and Cultural Research; National Institute of Culture and History.
170
Culleton, B.J., Prufer, K.M., Kennett, D.J., 2012. A Bayesian AMS 14C Chronology of the Classic Maya Center of Uxbenká, Belize. Journal of Archaeological Science 39:1572-1586.
Culleton, B.J., Prufer, K.M., Kennett, D.J., n.d. Changing Agricultural and Urban
Landscapes at the Classic Maya Center of Uxbenká, Belize. Manuscript in review, Geoarchaeology.
Curtis, J.H., Brenner, M., Hodell, D.A., Balser, R.A., Islebe, G.A., Hooghiemstra, H.,
1998. A Multi-Proxy Study of Holocene Environmental Change in the Maya Lowlands of Petén, Guatemala. Journal of Paleolimnology 19:139-159.
Curtis, J.H., Hodell, D.A, Brenner, M., 1996. Climate Variability on the Yucatán
Peninsula (Mexico) during the Past 3500 Years, and Implications for Maya Cultural Evolution. Quaternary Research 46:37-47.
Dean, W.E., Jr. 1974. Determination of Carbonate and Organic Matter in Calcareous
Sediments by Loss on Ignition: Comparison with Other Methods. Journal of Sedimentary Petrology 44:242-248.
Deevey, E.S., Rice, D.S., Rice, P.M., Vaughan, H.H., Brenner, M., Flannery, M.S., 1979.
Mayan Urbanism: Impact on a Tropical Karst Environment. Science 206:298-306. Demarest, A.A., 2004. Ancient Maya: The Rise and Fall of a Rainforest Civilization.
Cambridge University Press, Cambridge. Demarest, A.A., Rice, P., Rice, D. (eds.), 2004. The Terminal Classic in the Maya
Lowlands. University Press of Colorado, Boulder. Demarest, A.A., 2006. The Petexbatun Regional Archaeological Project: A
Multidisciplinary Study of the Maya Collapse. Vanderbilt Institute of Mesoamerican Archaeology, Nashville.
deMenocal, P.B., 2001 Cultural Responses to Climate Change during the Late Holocene.
Science 292:667-673. Denevan, W.M., 1992. The Pristine Myth: The Landscape of the Americas in 1492.
Annals of the Association of American Geographers 82:369-385. Diamond, J., Bellwood, P., 2003. Farmers and Their Languages: The First Expansions.
Science 300:597-603. Drennan, R.D., 2010. Statistics for Archaeologists. Springer. (2nd Edition) Dumond, D.E., 1975. The Limitation of Human Population: A Natural History. Science
187:713-721
171
Dunham, P.S., 1996. Resource Exploitation and Exchange among the Classic Maya: Some Initial Findings of the Maya Mountains Archaeological Project. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick, (ed.), pp. 315-334, University of Utah Press, Salt Lake City.
Dunham, P.S., Prufer, K.M., 1998. En la Cumbre del Clásico: Descubrimientos Recientes
en las Montañas Mayas en el Sur de Belice. In XI Simposio de Investigationes Arqueologicas en Guatemala, J.P. Laporte & H.L. Escobedo (eds.), pp. 165-170. Ministerio de Cultura y Deportes. Instituto de Antropología e Historia y Association Tikal, Guatemala City.
Dunning, N.P., 1996. A Reexamination of Regional Variability in the Pre-Hispanic
Agricultural Landscape. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 53-68. University of Utah Press, Salt Lake City.
Dunning, N.P., Beach, T., 2000. Stability and Instability in Prehispanic Maya
Landscapes. In Imperfect Balance: Landscape Transformation in the Precolumbian Americas. D.L. Lentz (ed.), pp.179-202. Columbia University Press, New York.
Dunning, N.P., Luzzadder-Beach, S., Beach, T., Jones, J., Scarborough, V., Culbert, T.,
2002. Arising from the Bajos: The Evolution of a Neotropical Landscape and the Rise of Maya Civilization. Annals of the Association of American Geographers 92:267-283.
Dunning, N.P., Rue, D., Beach, T., 1991. Ecología y Patrón de Asentamiento en la
Región de Petexbatún: Resultos Preliminares de la Temporada 1991. In Proyecto Arqueológico Regional Petexbatún, Informe No. 3, Tercera Temporada 1991. A.A. Demarest (ed.), pp. 829-847. Unpublished manuscript. Department of Anthropology, Vanderbilt University, Nashville.
Dunning, N.P., Rue, D., Beach, T., Covich, A., Traverse, A., 1998. Human-Environment
Interactions in a Tropical Watershed: The Paleoecology of Laguna Tamarindito, El Petén, Guatemala. Journal of Field Archaeology 25:139-151.
Dunning, N.P., Scarborough, V.L., Valdez, F. Jr., Luzzadder-Beach, S., Beach, T., Jones,
J.G., 1999. Temple Mountains, Sacred Lakes, and Fertile Fields: Ancient Maya Landscapes in Northwestern Belize. Antiquity 73:650-660.
Earle, T., 1987. Chiefdoms in Archaeological and Ethnohistorical Context. Annual
Review of Anthropology 16: 279-308. Ebert, C.E., Richards, L., Mustain, C., Aquino, V., Trask,W., Noldner, L., Prufer, K.M.,
2010. Excavations in Group D. In Uxbenká Archaeological Project: Report of the 2009 Field Season in the Toledo District, Belize, Central America. K.M. Prufer (ed), pp. 104-126. Belize Institute of Archaeology; Belize Institute of Social and Cultural Research; National Institute of Culture and History. Belmopan, Belize.
172
Erlandson, J.M. 2001. The Archaeology of Aquatic Adaptations: Paradigms for a New Millennium. Journal of Archaeological Research 9:287-350.
Erlandson, J.M., Rick, T.C., Braje, T.J., Casperson, M., Culleton, B.J., Fulfrost, B.
Garcia, T., Guthrie, D., Jew, N., Kennett, D.J., Moss, M.L., Reeder, L., Skinner, C., Watts, J., Willis, L., 2011. Paleoindian Seafaring, Maritime Technologies, and Coastal Foraging on California’s Channel Islands. Science 331:1181-1185.
Erlandson, J.M., Rick, T.C., Kennett, D.J., Walker, P.L., 2001. Dates, Demography, and
Disease: Cultural Contacts and Possible Evidence for Old World Epidemics among the Protohistoric Island Chumash. Pacific Coast Archaeological Society Quarterly 37:11-26.
FAO/WHO/UNU, 1991. Energy and Protein Requirements: Report of a Joint
FAO/WHO/UNU Expert Consultation. World Health Organization Technical Report Series 724. World Health Organization, Geneva. Avavilbe online: http://www.fao.org/docrep/003/AA040E/AA040E00.htm
Fash, W.L., Sharer, R.J., 1991. Sociopolitical Developments and Methodological Issues
at Copán, Honduras: A Conjunctive Perspective. Latin American Antiquity 2:166-187. Fedick, S.L. (ed.), 1996a. The Managed Mosaic: Ancient Maya Agriculture and Resource
Use. University of Utah Press, Salt Lake City. Fedick, S.L., 1996b. An Interpretive Kaleidoscope: Alternative Perspectives on Ancient
Agricultural Landscapes of the Maya Lowlands. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 107-131. University of Utah Press, Salt Lake City.
Fedick, S.L., 1996c. Conclusion: Landscape Approaches to the Study of Ancient Maya
Agriculture and Resource Use. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 335-347. University of Utah Press, Salt Lake City.
Fedick, S.L., Ford, A., 1990. The Prehistoric Agricultural Landscape of the Central Maya
Lowlands: An Examination of Local Variability in a Regional Context. World Archaeology 22:18-33.
Fitzhugh, B., Kennett, D.J., 2010. Seafaring Intensity and Island–Mainland Interaction
along the Pacific Coast of North America. In Global Origins and Development of Seafaring. A.Anderson, J.H. Barrett, K.V. Boyle (eds.), pp. 69-80. McDonald Institute for Archaeological Research, Cambridge.
Flannery, K.V., 1972. The Cultural Evolution of Civilizations. Annual Review of Ecology
and Systematics 3:399-426.
173
Folan, W.S., Marcus, J., Pincemin, S., del Rosario Dominguez Carrasco, M., Fletcher, L., Morales Lopez, A., 1995. Calakmul: New Data from an Ancient Maya Capitol in Campeche, Mexico. Latin American Antiquity 6:310–334.
Ford, A., 2005. Human Impacts on the Maya Forest: Linking the Past with the Present for
the Future of El Pilar. BRASS/El Pilar Program. Report on the 2004 field season. MesoAmerican Research Center, University of California, Santa Barbara, California, USA. http://marc.ucsb.edu/elpilar/chronicles/Field_Reports.phpi
Freidel, D.A., 1986. Archaeology at Cerros, Belize, Central America. Southern Methodist
University Press, Dallas, TX. Freidel, D.A., Schele, L., 1988. Kingship in the Late Preclassic Maya Lowlands: The
Instruments and Places of Ritual Power. American Anthropologist 90:547-567. Fretwell, S.D., 1972. Populations in a Seasonal Environment. Princeton University Press,
Princeton, NJ. Fretwell, S.D., Lucas Jr., H.L., 1969. On Territorial Behavior and Other Factors
Influencing Habitat Distribution in Birds. I. Theoretical Development. Acta Biotheoretica 19:16–36.
Garber, J.F., Brown, M.K., Awe, J.J., Hartman, C.J., 2004. Middle Formative Prehistory
of the Central Belize Valley: Examination of Architecture, Material Culture, and Sociopolitical Change at Blackman Eddy. In The Ancient Maya of the Belize Valley: Half a Century of Archaeological Research, J.F. Garber (ed.), pp. 25-47. University Press of Florida, Gainesville.
Gill, R.B., 2000. The Great Maya Droughts: Water, Life, and Death. University of New
Mexico Press, Albuquerque. Goebel, T., Waters, M.R., O’Rourke, D.H., 2008. The Late Pleistocene Dispersal of
Modern Humans in the Americas. Science 319:1497-1502. Graham, E.A., 1987. Resource diversity in Belize and its implications for models of
lowland trade. American Antiquity 52:753-767 Graham, E.A., 1994. Highlands of the Lowlands: Environment and Archaeology in the
Stann Creek District, Belize, Central America. Prehistory Press, Madison, WI. Grimm, N.B., Grove, J.M., Pickett, S.T.A., Redman, C.L., 2000. Integrated Approaches
to Long-Term Studies of Urban Ecological Systems. BioScience 50:571-584. Grube, N., MacLeod, B., Wanyerka, P., 1999. A Commentary on the Hieroglyphic
Inscriptions of Nim Li Punit, Belize. Research Reports on Ancient Maya Writing, No. 41. Center for Maya Research, Washington, DC.
174
Gurdjieff, G.I., 1973. Beelzebub’s Tales to His Grandson: An Objectively Impartial Criticism of the Life of Man. First Book. E.P. Dutton and Co., Inc., New York.
Hammond, N., 1975. Lubaantun: A Classic Maya Realm, Monographs 2. Harvard
University, Peabody Museum of Archaeology and Ethnology. Cambridge, MA Hammond, N., 1978. Cacao and Cobañeros: An Overland Trade Route between the Maya
High- lands and Lowlands. In Mesoamerican Communication Routes and Cultural Contacts. T.A. Lee, Jr. & C. Navarrete (eds.), pp. 19-26. New World Archaeological Foundation Papers, No. 40.
Hammond, N. 1991. Cuello: An Early Maya Community in Belize. Cambridge University
Press, Cambridge, UK. Hammond, N., Gerhardt, J.C., 1990. Early Maya Architectural Innovation at Cuello,
Belize. World Archaeology 21:461–481. Hammond, N., Howarth, S., Wilk, R.R., 1999. The Discovery, Exploration, and
Monuments of Nim Li Punit, Belize. Research Reports on Ancient Maya Writing, No. 40. Center for Maya Research, Washington, D.C.
Hammond, N., Wilk, R.R., Kosakowsky, L.J., 1991. Stratigraphy and Chronology in the
Reconstruction of Preclassic Development at Cuello. In Cuello: An Early Maya Community in Belize. N. Hammond (ed.), pp 23-69. Cambridge University Press, Cambridge, UK.
Hansen, R.D., 1994. Las Dinámicas Culturales y Ambientales de los Orígenes Mayas:
Estudios Recientes del Sitio Arqueológico Nakbe. In VII Simposio Arqueológico de Guatemala, 1993. J.L. Laporte & H.L. Escobedo (eds.), pp.115-122. Museo Nacional de Arqueología y Etnología. Guatemala City, Guatemala.
Hansen, R.D., 1995. Early Environmental Impact: The Ecological Consequences of
Incipient Maya Settlement. Technical report to the National Geographic Society, NGS 4984-93. Washington, D.C.
Hansen, R.D., 1998. Continuity and Disjunction: The Pre-Classic Antecedents of Classic
Maya Architecture. In Function and Meaning in Classic Maya Architecture. S.D. Houston (ed.), pp.49-122. Dumbarton Oaks Research Library and Collection.
Hansen, R.D., 2006. The First Cities: The Beginnings of Urbanization and State
Formation in the Maya Lowlands. In Maya: Divine Kings of the Rainforest, N. Grube (ed). pp. 50–65. Köneman, Cologne, Germany.
175
Harrison, P.D., 1993. Aspects of Water Management in the Southern Maya Lowlands. In Economic Aspects of Water Management in the Prehispanic New World. V.L. Scarborough & B.L. Isaac (eds.), pp. 71-119. Research in Economic Anthropology Supplement no. 7. JAI Press, Greenwich, CT.
Harrison, P.D., 1996. Settlement and Land Use in the Pulltrouser Swamp Archaeological
Zone, Northern Belize. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 177-190. University of Utah Press, Salt Lake City.
Harrison, P.D., 2006. Maya Architecture at Tikal. In Maya: Divine Kings of the
Rainforest, N. Grube (ed.), pp. 218–231. Köneman, Cologne, Germany. Harrison, P.D., Turner, B.L. II, (eds.), 1978. Pre-Hispanic Maya Agriculture. University
of New Mexico Press, Albuquerque. Hartshorn, G.S., Nicolait, L., Hartshorn, L., Bevier, G., Brightman, R., Cal, J., Cawich,
A., Davidson, W., Dubois, R., Dyer, C., Gibson, J., Hawley, W., Leonard, J., Nicolait, R., Weyer, D., White, H., Wright, C., 1984. Belize Country Environmental Profile: A Field Study. USAID Contract No. 505-0000-C-00- 3001-00. San José, Costa Rica: Trejos Hermanos Sucesores, S.A.
Haug, G.H., Gunther, D., Peterson, L.C., Sigman, D.M., Hughen, K.A., Aeschlimann, B.,
2003. Climate and the Collapse of Maya Civilization. Science 299:1731-1735. Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C., Röhl, U., 2001. Southward
Migration of the Intertropical Convergence Zone through the Holocene. Science 293:1304-1308.
Haviland, W.A., 1969. A New Population Estimate for Tikal, Guatemala. American
Antiquity 34: 429-433. Haviland, W.A., 1972. Family Size, Prehistoric Population Estimates, and the Ancient
Maya. American Antiquity 37:135-139. Healy, P.F., 2006. Preclassic Maya of the Belize Valley: Key Issues and Questions.
Research Reports in Belizean Archaeology 3:13-30. The Institute of Archaeology, NICH, Belize.
Healy, P.F., Lambert, J.D.H., Arnason, J.T., Hebda, R.J., 1983. Caracol, Belize: Evidence
of Ancient Maya Agricultural Terraces. Journal of Field Archaeology 10:397-410. Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss On Ignition as a Method for Estimating
Organic and Carbonate Content in Sediments: Reproducibility and Comparability of Results. Journal of Paleolimnology 25: 101–110.
176
Henderson, J., 1992. Variations on a Theme: A Frontier View of Maya Civilization. In New Theories on the Ancient Maya. E.C. Danien and R.J. Sharer (eds.), pp. 161-171. University of Pennsylvania Museum, Philadelphia.
Heyman, W.D., Kjerfve, B., 1999. Hydrological and Oceanographic Considerations for
Integrated Coastal Zone Management in Southern Belize. Environmental Management 24:229–245.
Hodell, D.A., Brenner, M., Curtis, J.H., 2000. Climate Change in the Northern American
Tropics and Subtropics since the Last Ice Age. In Imperfect Balance: Landscape Transformation in the Precolumbian Americas. D.L. Lentz (ed.), pp. 14-38. Columbia University Press, New York.
Hodell, D.A., Brenner, M., Curtis, J.H., 2005. Terminal Classic Drought in the Northern
Maya Lowlands Inferred from Multiple Sediment Cores in Lake Chichancanab (Mexico). Quaternary Science Reviews 24:1413-1427.
Hodell, D.A., Brenner, M., Curtis, J.H., Guilderson, T., 2001. Solar Forcing of Drought
Frequency in the Maya Lowlands. Science 292:1367-1370. Hodell, D.A., Curtis, J.H., Brenner, M., 1995. Possible Role of Climate in the Collapse of
Classic Maya Civilization. Nature 375:391-394. Islebe, G.A., Hooghiemstra, H., Brenner, M., Curtis, J.H., Hodell, D.A., 1996. A
Holocene Vegetation History from Lowland Guatemala. The Holocene 6:265-271. Johnson, A.W, Earle, T., 1987. The Evolution of Human Societies from Foraging Group
to Agrarian State. Stanford University Press, Stanford, CA. Jones, J.G., 1994. Pollen Evidence for Early Settlement and Agriculture in Northern
Belize. Palynology 18:207-213. Joyce, A.A., Mueller, R., 1992. The Social Impact of Anthropogenic Landscape
Modification in the Rıó Verde Drainage Basin, Oaxaca, Mexico. Geoarchaeology 7: 503–526.
Joyce, T.A., 1929. Report on the British Museum Expedition to British Honduras, 1929.
Journal of the Royal Anthropological Institute 59:439-459 Joyce, T.A., Cooper Clark, J., Thompson, J.E., 1927. Report on the British Museum
Expedition to British Honduras,1927. Journal of the Royal Anthropological Institute 57:297-323.
177
Keller, G., Stinnesbeck, W., Adatte, T., Holland, B., Stüben, D., Harting, M., De Leon, C, & De la Cruz, J., 2003. Spherule Deposits in Cretaceous–Tertiary Boundary Sediments in Belize and Guatemala. Journal of the Geological Society, London 160:783–795.
Kennett, D.J., 1998. Behavioral Ecology and the Evolution of Hunter-Gatherer Societies
on the Northen Channel Islands, California. Unpublished Doctoral Dissertation, Department of Anthropology. University of California, Santa Barbara.
Kennett, D.J., 2005. The Island Chumash: Behavioral Ecology of a Maritime Society.
University of California Press, Berkeley. Kennett, D.J., Anderson, A., Winterhalder, B., 2006. The Ideal Free Distribution, Food
Production, and the Colonization of Oceania. In: Behavioral Ecology and the Transition to Agriculture. D.J. Kennett & B. Winterhalder (eds.), pp. 265–288. University of California Press, Berkeley.
Kennett, D.J., Clifford, R., 2004. Flexible Strategies for Resource Defense on the
Northern Channel Islands of California: An Agent-Based Model. In The Archaeology of Insularity: Examining the Past in Island Environments. S. Fitzpatrick (ed.), pp. 21-50. Greenwood Press, West Port, CT.
Kennett, D.J., Culleton, B.J., Voorhies, B., Southon, J.R., 2011. Bayesian Analysis of
High Precision AMS 14C Dates from a Prehistoric Mexican Shellmound. Radiocarbon 53:245–259.
Kennett, D.J,. Ingram, B.L., Southon, J.R., Wise, K., 2002. Differences in 14C Age
Between Stratigraphically Associated Charcoal and Marine Shell from the Archaic Period Site of Kilometer 4, Southern Peru: Old Wood or Old Water? Radiocarbon 44:53-58.
Kennett, D.J,. Kennett, J.P., 2006. Sea Levels, Shorelines, Climate Change, and Cultural
Evolution in Southern Mesopotamia. Journal of Island & Coastal Archaeology 1:39-71.
Kennett, D.J., Piperno, D.R., Jones, J.G., Neff, H., Voorhies, B., Walsh, M.K., Culleton,
B.J., 2010. Pre-Pottery Farmers on the Pacific Coast of Southern Mexico. Journal of Archaeological Science 37:3401-3411.
Kennett, D.J., Winterhalder, B. (eds.), 2006. Behavioral Ecology and the Transition to
Agriculture. University of California Press, Berkeley.
178
Kennett, D.J., Winterhalder, B., 2008. Demographic Expansion, Despotism, and the Colonisation of East and South Polynesia. In Islands of Inquiry: Colonisation, Seafaring and the Archaeology of Maritime Landscapes (Terra Australis 29). G. Clark, F. Leach,& S. O’Connor (eds.), pp. 87–96. Australia National University Press, Canberra.
Kennett, D.J., Winterhalder, B., Bartruff, J., Erlandson, J.M., 2009. An Ecological Model
for the Emergence of Institutionalized Social Hierarchies on California’s Northern Channel Islands. In Pattern and Process in Cultural Evolution. S. Shennan (ed.), pp. 297–314. University of California Press, Berkeley.
Kirch, P.V., 2000. On the Road of the Winds: An Archaeological History of the Pacific
Islands before European Contact. University of California Press, Berkeley. Kirch, P.V., Hunt, T.L. (eds.),1997. Historical Ecology in the Pacific Islands: Prehistoric
Environmental and Landscape Change. Yale University Press, New Haven, CT. Klein, R.G., 1992. The Impact of Early People on the Environment: The Case of Large
Mammal Extinctions. In Human Impact on the Environment: Ancient Roots, Current Challenges. J.E. Jacobsen & J. Firor (eds.), pp. 13-34. Westview Press, Boulder, CO.
Kohler, T.A., van der Leeuw, S.E., 2007. The Model-Based Archaeology of Socionatural
Systems. School of American Research, Santa Fé, NM. Kolata, A.L., 1986. The Agricultural Foundations of the Tiwanaku State: A View from
the Heartland. American Antiquity 51:748-762. Laporte, J.P., 1994. Ixtonton, Dolores, Petén: Entidad Política del Noroeste de las
Montañas Mayas. Atlas Arquelógico de Guatemala, no. 2. Guatemala: Ministerio de Cultura y Deportes, Instituto de Antropologia e Historia
Laporte, J.P., 1996a. El Concepto de Entidad Segmentaria en la Historia Arqueológica
del Noroeste de las Montañas Mayas. Mayab 10: 25-32. Laporte, J.P., 1996b. La Region del Sureste de Petén, Guatemala, en la Arqueología de
las Tierras Bajas Centrales. In Arqueología Mesoamericana: Homenajea William T. Sanders, Vol. 2, pp.137-168, Instituto Nacional de Antropología y Historia, Mexico.
Laporte, J.P., 2001. La Gloria-Sacul, Petén: Un Sitio del Preclasico en las Montañas
Mayas de Guatemala. Mayab 14:17-29. Laporte, J.P., Ramos, C.E., 1998. Sacul, Dolores, Petén: Excavation, Arquitectura y
Hallazgos. Utz’ib 2(4):1-22.
179
LeCount, L.J., Yaeger, J., Leventhal, R.M., Ashmore, W., 2002. Dating the Rise and Fall of Xunantunich, Belize: A Late and Terminal Classic Lowland Maya Regional Center. Ancient Mesoamerica 13:41-63.
Lee, C.T., Tuljapurkar, S., Vitousek, P., 2008. Population and Prehistory I: Food-
Dependent Population Growth in Constant Environments. Theoretical Population Biology 73:473–482
Lee, C.T., Puleston, C., Tuljapurkar, S. 2009. Population and Prehistory III: Space-
Limited Human Populations in Constant Environments. Theoretical Population Biology 76:179-188.
Lentz, D.L., (ed.), 2000. Imperfect Balance: Landscape Transformation in the
Precolumbian Americas. Columbia University Press, New York. Lentz, D.L., Hockaday, B., 2009. Tikal Timbers and Temples: Ancient Maya
Agroforestry and the End of Time. Journal of Archaeological Science 36:1342-1353. Leventhal, R.M., 1990. Southern Belize: An Ancient Maya Region. In Vision and
Revision in Maya Studies. F.S. Clancy & P.D. Harrison (eds.), pp. 125-141. University of New Mexico Press, Albuquerque.
Leventhal, R.M., 1992. The Development of a Regional Tradition in Southern Belize. In
New Theories on the Ancient Maya. E.C. Danien & R.J. Sharer (eds.), pp. 145-154. The University Museum, University of Pennsylvania, Philadelphia.
Lohse, J.C., 2010. Archaic Origins of the Lowland Maya. Latin American Antiquity
21:312-352. Lohse, J.C., Awe, J., Griffith, C., Rosenswig, R.M., Valdez, F., Jr., 2006. Preceramic
Occupations in Belize: Updating the Paleoindian and Archaic record. Latin American Antiquity 17:209-226.
Lucero, L.J., 2002. The Collapse of the Classic Maya: A Case for the Role of Water
Control. American Anthropologist 104:814–826. MacIntyre, I.G., Toscano, M.A., Lighty, R.G., Bond, G.B., 2004. Holocene History of the
Mangrove Islands of Twin Cays, Belize, Central America. Atoll Research Bulletin No. 510. National Museum Of Natural History, Smithsonian Institution, Washington, D.C.
Manning, S.W., Bronk Ramsey, C., Kutschera, W., Higham, T., Kromer, B., Steier, P.,
Wild, E.M., 2006. Chronology for the Aegean Late Bronze Age 1700−1400 BC. Science 312:565−9.
180
Marcus, J., 1993. Ancient Maya Political Organization. In Lowland Maya Civilization in the Eighth Century A.D. J.A. Sabloff & J.S. Henderson (eds.), pp. 111-183. Dumbarton Oaks, Washington, D.C.
Marcus, J., 2003. The Maya and Teotihuacan. In The Maya and Teotihuacan:
Reinterpreting Early Classic Interaction. G.E. Braswell (ed.), pp. 337-356. University of Texas Press, Austin.
Martin, P.S., 2005. Twilight of the Mammoths: Ice Age Extinctions and the Rewilding of
America. University of California Press, Berkeley, CA. Martin, S., Grube, N., 2008. Chronicle of the Maya Kings and Queens: Deciphering the
Dynasties of the Ancient Maya. Thames and Hudson Ltd., London. Masson, M.A., Freidel, D.A., 2002. Ancient Maya Political Economies. AltaMira Press,
Walnut Creek, CA. McClure, S.B., Jochim, M.A., Barton, C.M., 2006. Human Behavioral Ecology,
Domestic Animals, and Land Use During the Transition to Agriculture in Valencia, Eastern Spain. In Behavioral Ecology and the Transition to Agriculture. D.J. Kennett & B. Winterhalder (eds.), pp. 197-217. University of California Press, Berkeley.
McKillop, H., 1989. Coastal Maya Trade: Obsidian Densities from Wild Cane Cay,
Belize. In Prehistoric Maya Economies of Belize. P.A. McAnany & B.L. Isaac (eds.), pp. 17-56. Research in Economic Anthropology, Supplement 4. JAI Press, Greenwich, CT.
McKillop, H., 1995. Underwater Archaeology, Salt Production, and Coastal Maya Trade
at Stingray Lagoon, Belize. Latin American Antiquity 6:214-228. McKillop, H., 1996. Ancient Maya Trading Ports and the Integration of Long-Distance
and Regional Economies: Wild Cane Cay in South-Coastal Belize. Ancient Mesoamerica 7:49-62
McKillop, H., 2005a. Finds in Belize Document Late Classic Maya Salt Making and
Canoe Transport. Proceedings of the National Academy of Sciences 102:5630-5634. McKillop, H., 2005b. In Search of Maya Sea Traders. Texas A&M University Press,
College Station, TX. McKillop, H., 2006. The Ancient Maya: New Perspectives, W.W. Norton, New York. McKillop, H., 2008. Salt: White Gold of the Ancient Maya. University of Florida Press,
Gainesville, FL.
181
McKillop, H., 2010. Ancient Maya Canoe Navigation and Its Implications for Classic to Postclassic Maya Economy and Sea Trade: A View from the South Coast of Belize. Journal of Caribbean Archaeology, Special Publication 3:93-105.
Miller, T.E., 1996. Geologic and Hydrologic Controls on Karst and Cave Development in
Belize. Journal of Cave and Karst Studies 58:100-120. Moyes, H., Awe, J.J., Brook, G.A., Webster, J.W., 2009. The Ancient Maya Drought
Cult: Late Classic Cave Use in Belize. Latin American Antiquity 20:175-206. Mueller, A.D., Islebe, G.A., Hillesheim, M.B., Grzesik, D.A., Anselmett, F.S., Ariztegui,
D., Brenner, M., Curtis, J.H., Hodell, D.A., Venz, K.A., 2009. Climate Drying and Associated Forest Decline in the Lowlands of Northern Guatemala during the Late Holocene. Quaternary Research 71:133–141.
Mueller, A.D., Islebe, G.A., Anselmetti, F.S., Ariztegui, D., Brenner, M., Hodell, D.A.,
Hajdas, I., Hamann, Y., Haug, G.H., Kennett, D.J., 2010. Recovery of the Forest Ecosystem in the Tropical Lowlands of Northern Guatemala after Disintegration of Classic Maya Polities. Geology 38:523–526.
Nations, J.D., 2006. The Maya Tropical Forest : People, Parks, and Ancient Cities.
Austin: University of Texas Press, Austin, TX. Nations, J.D., Nigh, R., 1980. Evolutionary Potential of Lacandon Maya Sustained-Yield
Tropical Forest Agriculture. Journal of Anthropological Research 36:1–30. Neff, H., Pearsall, D.M., Jones, J.G., Arroyo, B., Collins, S.K., Friedel, D.E., 2006. Early
Maya Adaptive Patterns: Mid-Late Holocene Paleoenvironmental Evidence from Pacific Guatemala. Latin American Antiquity 17:287-315.
Netting, R.McC., 1993. Smallholders, Householders. Stanford University Press. O’Hara, S.L., Street-Perrott, F.A., Burt, T.P., 1993. Accelerated Soil Erosion around a
Mexican Lake Caused by Prehispanic Agriculture. Nature 362:48-51. Pendergast, D.M., 1981. Lamanai, Belize: Summary of Excavation Results, 1974–1980.
Journal of Field Archaeology 8:29–53. Piperno, D.R., Pearsall, D.M., 1998. The Origins of Agriculture in the Lowland
Neotropics. Academic Press, San Diego. Pohl, M.D., Bloom, P., 1996. Prehistoric Maya Farming in the Wetlands of Northern
Belize: More Data from Albion Island and Beyond. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 145-164. University of Utah Press, Salt Lake City.
182
Pohl, M.D., Pope, K.O., Jones, J.G., Jacob, J.S., Piperno, D.R., deFrance, S.D., Lentz, D.L., Gifford, J.A., Danforth, M.E., Josserand, J.K., 1996. Early Agriculture in the Maya Lowlands. Latin American Antiquity 7:355-372.
Pope, K.O., Pohl, M.D., Jacob, J.S., 1996. Formation of Ancient Maya Wetland Fields:
Natural and Anthropogenic Processes. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 165-176. University of Utah Press, Salt Lake City.
Pope, K., Pohl, M., Jones, J., Lentz, D,. von Nagy, C., Vega, F., Quitmyer, I., 2001.
Origin and Environmental Setting of Ancient Agriculture in the Lowlands of Mesoamerica. Science 292:1370.
Price, T.D., Gebauer, A.B. (eds.), 1995. Last Hunters-First Foragers: New Perspectives
on the Prehistoric Transition to Agriculture. School of American Research Press, Santa Fe, New Mexico.
Prufer, K.M., 2005. The Early Classic in Southern Belize: A Regional View From
Uxbenka and Ek Xux. Research Reports in Belizean Archaeology 2:169-178. Prufer, K.M., Hurst, W.J., 2007. Chocolate and the Underworld Space of Death: The
Recovery of Intact Cacao from an Early Classic Mortuary Cave. Ethnohistory 54:273-303.
Prufer, K.M., Moyes, H., Culleton, B.J., Kindon, A., Kennett, D.J. 2011. Formation of a
Complex Polity on the Eastern Periphery of the Maya Lowlands. Latin American Antiquity 22:199-223.
Prufer, K.M., Wanyerka, P.J., 2005. A New Early Classic Stela from Uxbenká, Belize.
Mexicon xxiii:102-103. Puleston, C., Tuljapurkar, S., 2008. Population and Prehistory II: Space-Limited Human
Populations in Constant Environments. Theoretical Population Biology 74:147-160. Puleston, D.E., 1978. Terracing, Raised Fields, and Tree-Cropping in the Maya
Lowlands: A New Perspective on the Geography of Power. In Pre-Hispanic Maya Agriculture. P.D. Harrison & B.L. Turner II (eds.), pp. 235-245. University of New Mexico Press, Albuquerque.
Puleston, D.E., 1982. The Role of Ramon in Maya Subsistence. In Maya Subsistence:
Studies in Memory of Dennis E. Puleston. K.V. Flannery (ed.), pp.353-366. Academic Press, New York.
Redman, C.L., 1992. The Impact of Food Production: Short-Term Strategies and Long-
Term Consequences. In Human Impact on the Environment: Ancient Roots, Current Challenges. J. Jacobsen & J. Firor (eds.), pp. 35-49. Westview Press, Boulder, CO.
183
Redman, C.L., 1999. Human Impacts on Ancient Environments. University of Arizona Press, Tucson.
Redman, C.L., 2005. Resilience Theory in Archaeology. American Anthropologist
107:70–77. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.J., Bronk
Ramsey, C., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., & Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 Radiocarbon Age Calibration Curves, 0–50,000 years cal BP. Radiocarbon 51:1111-1150.
Reina, R.E., 1967. Milpas and Milperos: Implications for Prehistoric Times. American
Anthropologist 69:1-20. Reina, R.E., Hill, R.M., 1980. Lowland Maya Subsistence: Notes from Ethnohistory and
Ethnography American Antiquity 45:74-79. Reith, M., Aquino, V., Trask, W., Merriman, C.W., Kalosky, E.K., Noldner, L., 2011.
2010 Site Core Excavations at Uxbenká – Group I And Group A. In Uxbenká Archaeological Project: Report of the 2010 Field Season. K.M. Prufer, D.J. Kennett & E. Kalosky (eds.), pp. 19-48. Submitted to: Belize Institute of Archaeology; Belize Institute of Social and Cultural Research; National Institute of Culture and History.
Rice, D.S., 1993. Eighth-Century Physical Geography, Environment, and Natural
Resources in the Maya Lowlands. In Lowland Maya Civilization in the Eighth Century A.D. J.A. Sabloff & J.A. Henderson (eds.), pp.11-63. Dumbarton Oaks, Washington, D.C.
Rice, D.S., Culbert, T.P., 1990. Historical Contexts for Population Reconstruction in the
Maya Lowlands. In Precolumbian Population History in the Maya Lowlands. T.P. Culbert & D.S. Rice (eds.), pp. 1-36. University of New Mexico Press, Albuquerque.
Rick, J.W., 1987. Dates as Data: An Examination of the Peruvian Preceramic
Radiocarbon Record. American Antiquity 52:55-73 Rosenmeier, M.F., Hodell, D.A., Brenner, M., Curtis, J.H., Guilderson, T.P., 2002a. A
4000-year Lacustrine Record of Environmental Change in the Southern Maya Lowlands, Petén, Guatemala. Quaternary Research 57:183-190.
Rosenmeier, M.F., Hodell, D.A., Brenner, M., Curtis, J.H., Martin, J.B., Anselmetti, F.S.,
Ariztegui, D., Guilderson, T.P., 2002b. Influence of Vegetation Change on Watershed Hydrology: Implications for Paleoclimatic Interpretation of Lacustrine δ18O Records. Journal of Paleolimnology 27:117-131.
184
Rosenswig, R.M., Kennett, D.J., 2008. Reassessing San Estevan’s Role in the Late Formative Political Landscape of Northern Belize. Latin American Antiquity 19:124-146.
Rosenswig, R.M., 2009. Early Mesoamerican Garbage: Ceramic and Daub Discard
Patterns from Cuauhtémoc, Soconusco, Mexico. Journal of Archaeological Method Theory 16:1–32.
Ross, N.J., 2011. Modern Tree Species Composition Reflects Ancient Maya “Forest
Gardens” in Northwest Belize. Ecological Applications 21:75-84. Ross, N.J., Rangel, T.F., 2011. Ancient Maya Agroforestry Echoing through Spatial
Relationships in the NW Belize. Biotropica 43:141-148. Ruxton, G.D., 2006. The Unequal Variance t-Test is an Underused Alternative to
Student’s t-Test and the Mann–Whitney U-Test. Behavioral Ecology doi:10.1093/beheco/ark016.
Saturno, W. A., Stuart, D., Beltran, B., 2006. Early Maya Writing at San Bartólo,
Guatemala. Science 311:1281-1283. Scarborough, V.L., 1983. A Preclassic Maya Water System. American Antiquity 48:720–
744. Scarborough, V.L., 2003. The Flow of Power: Ancient Water Systems and Landscapes.
School of American Research Press, Santa Fe. Schele, L., Freidel, D.A., 1990. A Forest of Kings. Morrow, New York. Schele, L., Looper, M., 1996. Notebook for the XXth Maya Hieroglyphic Forum, March
9-10, 1996. Department of Art and Art History, University of Texas at Austin. Schiffer M.B., 1986. Radiocarbon Dating and the “Old Wood” Problem: The Case of the
Hohokam Chronology. Journal of Archaeological Science 13:13-30. Schiffer, M.B., 1987. Formation Processes of the Archaeological Record. University of
Utah Press, Salt Lake City. Schortman, E., Urban, P., 1994. Living on the Edge: Core/Periphery Relations in Ancient
Southeastern Mesoamerica. Current Anthropology 35:401-430. Sharer, R., 2003. Tikal and Copán Dynastic Founding. In Tikal: Dynasties, Foreigners,
and Affairs of State. J.A. Sabloff (ed.), pp. 319-353. School of American Research Press, Santa Fe.
185
Shennan, S., 1997. Quantifying Archaeology. Edinburgh University Press. Shennan, S., Edinborough, K., 2007. Prehistoric Population History: From the Late
Glacial to the Late Neolithic in Central and Northern Europe. Journal of Archaeological Science 34:1339–1345.
Shott, M.J., 2006. Formation Theory’s Past and Future: Introduction to the Volume. In
Formation Theory in Archaeology: Readings from American Antiquity and Latin American Antiquity. M. Shott (ed.), pp. 1-16. The SAA Press, Washington, DC.
Smith, E.A., 2000. Three Styles of Evolutionary Analysis of Human Behavior. In
Adaptation and Human Behavior: An Anthropological Perspective. L. Cronk, N. Chagnon, & W. Irons (eds.), pp.27-46. Aldine de Gruyter, Hawthorne, NY.
Smith, E.A., Winterhalder, B., 1992. Natural Selection and Decision Making: Some
Fundamental Principles. In Evolutionary Ecology and Human Behavior. E.A. Smith & B. Winterhalder (eds.), pp.25-60. Aldine de Gruyter, Hawthorne, NY.
Smith, B.D., 1998. The Emergence of Agriculture. Scientific American Library, New
York. Smith, B.D., 2001. Low-Level Food Production. Journal of Archaeological Research
9:1-43. Smith, B.D., 2007. The Ultimate Ecosystem Engineers. Science 315:1797-1798. Stadelman, R., 1940. Maize Cultivation in Northwestern Guatemala. Contributions to
American Anthropology and History 6. Carnegie Institution of Washington, Washington, D.C.
Steele, J., 2010. Radiocarbon Dates as Data: Quantitative Strategies for Estimating
Colonization Front Speeds and Event Densities. Journal of Archaeological Science 37:2017-2030.
Stahle, D.W., Villanueva Diaz, J., Burnette, D.J., Cerano Paredes, J., Heim, R.R. Jr., Fye,
F.K., Acuna Soto, R., Therrell, M.D., Cleaveland, M.K., Stahle, D.K., 2011. Major Mesoamerican Droughts of the Past Millennium. Geophysical Research Letters 38, L05703, doi:10.1029/2010GL046472.
Steier, P., Rom, W., 2000. The Use of Bayesian Statistics for 14C Dates of
Chronologically Ordered Samples: A Critical Analysis. Radiocarbon 42:183-198. Steinberg, M.K. 1998a. Mopan Maya Forest Resources in Southern Belize. Geographical
Review 88:131-137
186
Steinberg, M.K. 1998b. Political Ecology and Cultural Change: Impacts on Swidden-Fallow Agroforestry Practices among the Mopan Maya in Southern Belize. Professional Geographer 50:407-417.
Strahler, A.N., 1957. Quantitative Analysis of Watershed Geomorphology. Transactions
of the American Geophysical Union 38:913-920. Stuiver, M., and Polach, H.A., 1977. Discussion: Reporting of 14C data, Radiocarbon
19:355-363. Sutherland, W.J. 1996. From Individual Behaviour to Population Ecology. Oxford
University Press, Oxford. Tax, S., 1963. Penny Capitalism: A Guatemalan Indian Economy. University of Chicago
Press, Chicago, IL. Thomas, D.H., 1986. Refiguring Anthropology: Probability and Statistics. Waveland
Press. Toledo Maya Cultural Council (TMCC), 1997. Maya Atlas: The Struggle to Preserve
Maya Land in Southern Belize. North Atlantic Books, Berkeley. Trigger, B.G., 2003. Understanding Early Civilizations: A Comparative Study. Cambridge
University Press, Cambridge. Tuljapurkar, S., Lee, C. Figgs, M., 2007. Demography and Food in Early Polynesia. In
The Growth and Collapse of Pacific Island Societies. P.V. Kirch & J. Rallu (eds.), pp. 35-51. University of Hawaii Press, Manoa.
Turchin, P., 2003. Historical Dynamics: Why States Rise and Fall. Princeton University
Press, NJ. Turner, B.L., II, 2010. Unlocking the Ancient Maya and their Environment: Paleo-
evidence and Dating Resolution. Geology 38:575-576. Turner, B.L., II, Harrison, P.D., (eds.), 1983. Pulltrouser Swamp: Ancient Maya Habitat,
Agriculture and Settlement in Northern Belize. University of Texas Press, Austin. University of Arkansas Division of Agriculture, 2006. Equilibrium Moisture Content
Table Builder. Available online: http://www.aragriculture.org/storage_drying/default.htm
van der Leeuw, S., Redman, C.L., 2002. Placing Archaeology at the Center of Socio-
natural Studies. American Antiquity 67: 597-605.
187
Wainwright, J.D., 2007. First Affidavit of Joel David Wainwright. Document of the Supreme Court of Belize.
Wanyerka, P.J., 2004. The Southern Belize Epigraphic Project: The Hieroglyphic
Inscriptions of Southern Belize. Report submitted to the Foundation for the Advancement of Mesoamerican Studies, Incorporated. Available at: http://www.famsi.org/reports/00077/index.html.
Wanyerka, P.J., 2009. Classic Maya Political Organization: Epigraphic Evidence of
Hierarchical Organization in the Southern Maya Mountains Region of Belize. Unpublished Doctoral Dissertation, Department of Anthropology. Southern Illinois University, Carbondale, IL.
Webster D., 1985. Surplus, Labor, and Stress in Late Classic Maya Society. Journal of
Anthropological Research 41:375-399. Webster, D., 2002. The Fall of the Ancient Maya. Thames and Hudson, Ltd., London. Webster D., Freter, A., 1990. The Demography of Late Classic Copán. In Precolumbian
Population History in the Maya Lowlands. T.P. Culbert & D.S. Rice (eds.), pp. 37-62. University of New Mexico Press, Albuquerque.
Webster, D., Freter, A., Gonlin, N., 2000. Copán: The Rise and Fall of an Ancient Maya
Kingdom. Wadsworth, Belmont, CA. Webster, D., Freter, A., Storey, R., 2004. Dating Copán Culture History: Implications for
the Terminal Classic and the Collapse. In The Terminal Classic in the Maya Lowlands, A. Demarest, P. Rice, & D. Rice, (eds.), pp. 231-259. University Press of Colorado, Boulder, CO.
Webster, J.W., Brook, G.A., Railsback, L.B., Cheng, H., Edwards, R.L., Alexander, C.,
Reeder, P.P., 2007. Stalagmite Evidence from Belize Indicating Significant Droughts at the Time of Preclassic Abandonment, the Maya Hiatus, and the Classic Maya Collapse. Palaeogeography, Palaeoclimatology, Palaeoecology 250:1–17.
Weintraub, B., 1994. Geographica. National Geographic 185:4. Weiss, K.M., 1976. Demographic Theory and Anthropological Inference. Annual Review
of Anthropology 5:351-381. Wilk, R.R., 1984. Agriculture, Ecology, and Domestic Organization among the Kekchi
Maya. Unpublished Doctoral Dissertation, Department of Anthropology. University of Arizona, Tucson.
Wilk, R.R., 1991. Household Ecology: Economic Change and Domestic Life among the
Kekchi Maya in Belize. University of Arizona Press, Tucson.
188
Wingard, J.D., 1992. The Roll of Soils in the Development and Collapse of Classic Maya Civilization at Copán, Honduras. Unpublished Doctoral Dissertation, Department of Anthropology. The Pennsylvania State University, University Park.
Wingard, J.D., 1996. Interactions between Demographic Processes and Soil Resources in
the Copán Valley, Honduras. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use. S.L. Fedick (ed.), pp. 207-235. University of Utah Press, Salt Lake City.
Winterhalder, B., 1994. Concepts in Historical Ecology. A Perspective from Evolutionary
Ecology. In Historical Ecology: Cultural Knowledge and Changing Landscapes. C. Crumley (ed.), pp. 17-41. School of American Research Press, Santa Fe, NM.
Winterhalder, B., Baillargeon, W., Cappelletto, F., Daniel, I.R., Prescott, C., 1988 The
Population Ecology of Hunter-Gatherers and their Prey. Journal of Anthropological Archaeology 7:289-328.
Winterhalder, B., Goland, C., 1993. On Population, Foraging Efficiency, and Plant
Domestication. Current Anthropology 34:710-715. Winterhalder, B., Kennett, D.J., 2006. Behavioral Ecology and the Transition from
Hunting and Gathering to Agriculture. In: Behavioral Ecology and the Transition to Agriculture. D.J. Kennett & B. Winterhalder (eds.), pp. 1-21. University of California Press, Berkeley.
Winterhalder, B., Kennett, D.J., Grote, M.N., Bartruff, J., 2010. Ideal Free Settlement of
California’s Northern Channel Islands. Journal of Anthropological Archaeology 29: 469-490.
Winterhalder, B., Lu, F., 1997. A Forager-Resource Population Ecology Model and
Implications for Indigenous Conservation. Conservation Biology 11:1354-1362. Wolfman, D., 1990. Mesoamerican Chronology and Archaeomagnetic Dating, A.D. 1-
1200. In Archaeomagnetic Dating, J.L. Eighmy & R.S. Sternberg (eds.), pp. 261-308. University of Arizona Press, Tucson.
Wood, J.W., 1998. A Theory of Preindustrial Population Dynamics: Demography,
Economy, and Well-Being in Malthusian Systems. Current Anthropology 39:99-135. Wood, J.W., Milner, G.R., Harpending, H.C., Weiss, K.M., 1992. The Osteological
Paradox: Problems of Inferring Prehistoric Health from skeletal samples. Current Anthropology 33:343–70.
Wooller, M.J., Morgan, R., Fowell, S., Behling, H., Fogel, M., 2007. A Multiproxy Peat
Record of Holocene Mangrove Palaeoecology from Twin Cays, Belize. The Holocene 17:1129–1139.
189
Wright, A.C.S., Romney, D.H., Arbuckle, R.H., Vial, V.E., 1959. Land in British Honduras. Colonial Research Publications No. 24. Her Majesty’s Stationery Office, London.
Wyatt, A.R., 2008. Gardens on Hills: Ancient Maya Terracing and Agricultural
Production at Chan, Belize Valley. Unpublished Doctoral Dissertation, Department of Anthropology. University of Illinois at Chicago.
Zarger, R.K., 2002. Children’s Ethnoecological Knowledge: Situated Learning and the
Cultural Transmission of Subsistence Knowledge and Skills among Q’eqchi’ Maya. Unpublished Doctoral Dissertation, Department of Anthropology. University of Georgia, Athens, GA.
Zeder, M., 1991. Feeding Cities: Specialized Animal Economy in the Ancient Near East.
Smithsonian Series in Archaeological Inquiry. Smithsonian Institution Press, Washington, D.C.
Zimmerman Holt, J., 2009. Rethinking the Ramey State: Was Cahokia the Center of a
Theater State? American Antiquity 74:231-254
Top Related